SUMMARY TECHNICAL REPORT OF THE NATIONAL DEFENSE RESEARCH COMMITTEE ■is information affecting the national States within the meaning of the . C., 31 and 82, as amended. Its trans- m of its contentoin any manner to afr prohibited & |d RESTRICTED i||accordancje *with |the be- ll conjoin material RE- f(e Othm» chapters may or none. The n|ad$- is advised agencies listed,on the reverse in*nt classification of any material. Manuscript and illustrations for this volume were prepared for publication by the Summary Reports Group of the Columbia University Division of War Research under con- tract OEMsr-1131 with the Office of Scientific Research and Development. This volume was printed and bound by the Columbia University Press. Distribution of the Summary Technical Report of NDRC has been made by the War and Navy Departments. Inquiries concerning the availability and distribution of the Summary Technical Report volumes and microfilmed and other refer- ence material should be addressed to the War Department Library, Room 1A-522, The Pentagon, Washington 25, D, C., or to the Office of Naval Research, Navy Department, Atten- tion : Reports and Documents Section, Washington 25, D. C. Copy No. b-2 This volume, like the seventy others of the Summary Tech- nical Report of NDRC, has been written, edited, and printed under great pressure. Inevitably there are errors which have slipped past Division readers and proofreaders. There may be errors of fact not known at time of printing. The author has not been able to follow through his writing to the final page proof. Please report errors to: JOINT RESEARCH AND DEVELOPMENT BOARD PROGRAMS DIVISION (STR ERRATA) WASHINGTON 25, D. C. A master errata sheet will be compiled from these reports and sent to recipients of the volume. Your help will make this book more useful to other readers and will be of great value in preparing any revisions. SUMMARY TECHNICAL REPORT OF THE APPLIED PSYCHOLOGY PANEL, NDRC VOLUME 1 HUMAN FACTORS IN MILITARY EFFICIENCY APTITUDE AND CLASSIFICATION OFFICE OF SCIENTIFIC RESEARCH AND DEVELOPMENT VANNEVAR BUSH, DIRECTOR NATIONAL DEFENSE RESEARCH COMMITTEE JAMES B. CONANT, CHAIRMAN APPLIED PSYCHOLOGY PANEL CHARLES W. BRAY, CHIEF WASHINGTON, D. C., 1946 NATIONAL DEFENSE RESEARCH COMMITTEE James B. Conant, Chairman Richard C. Tolman, Vice Chairman Roger Adams Army Representative1 Frank B. Jewett Navy Representative2 Karl T. Compton Commissioner of Patents3 Irvin Stewart, Executive Secretary representatives in order of service: Maj. Gen. G. V. Strong Col. L. A. Denson Maj. Gen. R. C. Moore Col. P. R. Faymonville Maj. Gen. C. C. Williams Brig. Gen. E. A. Regnier Brig. Gen. W. A. Wood, Jr. Col. M. M. Irvine Col. E. A. Routheau 2Navy representatives in order of service: Rear Adm. H. G. Bowen Rear Adm. J. A. Purer Capt. Lybrand P. Smith Rear Adm. A. H. Van Keuren Commodore H. A. Schade 3Commissioners of Patents in order of service: Conway P. Coe Casper W. Ooms NOTES ON THE ORGANIZATION OF NDRC The duties of the National Defense Research Committee were (1) to recommend to the Director of OSRD suit- able projects and research programs on the instrumen- talities of warfare, together with contract facilities for carrying out these projects and programs, and (2) to administer the technical and scientific work of the con- tracts. More specifically, NDRC functioned by initiating research projects on requests from the Army or the Navy, or on requests from an allied government trans- mitted through the Liaison Office of OSRD, or on its own considered initiative as a result of the experience of its members. Proposals prepared by the Division, Panel, or Committee for research contracts for performance of the work involved in such projects were first reviewed by NDRC, and if approved, recommended to the Director of OSRD. Upon approval of a proposal by the Director, a contract permitting maximum flexibility of scientific effort was arranged. The business aspects of the con- tract, including such matters as materials, clearances, vouchers, patents, priorities, legal matters, and admin- istration of patent matters we£e handled by the Execu- tive Secretary of 0£>RD. \ A*-' ; Originally work through five divisions, each headed by one of the NDRC members. These were: Division A—Armor and Ordnance Division B—Bombs, Fuels, Gases, & Chemical Problems Division C—Communication and Transportation Division D—Detection, Controls, and Instruments Division E—Patents and Inventions In a reorganization in the fall of 1942, twenty-three administrative divisions, panels, or committees were created, each with a chief selected on the basis of his outstanding work in the particular field. The NDRC members then became a reviewing and advisory group to the Director of OSRD. The final organization was as follows: Division 1—Ballistic Research Division 2—Effects of Impact and Explosion Division 3—Rocket Ordnance Division 4—Ordnance Accessories Division 5—New Missiles Division 6—Sub-Surface Warfare Division 7—Fire Control Division 8—Explosives Division 9—Chemistry Division 10—Absorbents and Aerosols Division 11—Chemical Engineering Division 12—Transportation Division 13—Electrical Communication Division 14—Radar Division 15—Radio Coordination Division 16—Optics and Camouflage Division 17—Physics Division 18—War Metallurgy Division 19—Miscellaneous Applied Mathematics Panel Applied Psychology Panel Committee on Propagation Tropical Deterioration Administrative Committee RESTRICTED NDRC FOREWORD As events of the years preceding 1940 re- vealed more and more clearly the serious- ness of the world situation, many scientists in this country came to realize the need of organ- izing scientific research for service in a national emergency. Recommendations which they made to the White House were given careful and sym- pathetic attention, and as a result the National Defense Research Committee [NDRC] was formed by Executive Order of the President in the summer of 1940. The members of NDRC, appointed by the President, were instructed to supplement the work of the Army and the Navy in the development of the instrumentalities of war. A year later, upon the establishment of the Office of Scientific Research and Development [OSRD], NDRC became one of its units. The Summary Technical Report of NDRC is a conscientious effort on the part of NDRC to summarize and evaluate its work and to present it in a useful and permanent form. It comprises some seventy volumes broken into groups cor- responding to the NDRC Divisions, Panels, and Committees. The Summary Technical Report of each Divi- sion, Panel, or Committee is an integral survey of the work of that group. The first volume of each group’s report contains a summary of the report, stating the problems presented and the philosophy of attacking them, and summarizing the results of the research, development, and training activities undertaken. Some volumes may be “state of the art” treatises covering subjects to which various research groups have contributed information. Others may contain descriptions of devices developed in the labora- tories. A master index of all these divisional, panel, and committee reports which together constitute the Summary Technical Report of NDRC is contained in a separate volume, which also includes the index of a microfilm record of pertinent technical laboratory reports and ref- erence material. Some of the NDRC-gponsored researches which had been declassified by the end of 1945 were of sufficient popular interest that it was found desirable to report them in the form of monographs, such as the series on radar by Division 14 and the monograph on sampling inspection by the Applied Mathematics Panel. Since the material treated in them is not du- plicated in the Summary Technical Report of NDRC, the monographs are an important part of the story of these aspects of NDRC research. In contrast to the information on radar, which is of widespread interest and much of which is released to the public, the research on subsurface warfare is largely classified and is of general interest to a more restricted group. As a consequence, the report of Division 6 is found almost entirely in its Summary Technical Report, which runs to over twenty volumes. The extent of the work of a Division cannot therefore be judged solely by the number of volumes devoted to it in the Summary Technical Report of NDRC: account must be taken of the monographs and available reports published elsewhere. The Applied Psychology Panel, under the di- rection first of W, S. Hunter and later of C. W. Bray, comprised a small group of psychologists and personnel specialists whose responsibility was to aid in refining and standardizing Army and Navy personnel procedures. The Panel de- vised selection and classification tests; it devel- oped training methods; it improved the design of much equipment. The work of the Panel proved that it is as important to get the right man for a military job as it is to get the right ammunition for his gun. The achievements of the Applied Psychology Panel cannot be measured in quantitative terms. But one can, for example, estimate with cer- tainty that the tests devised to eliminate the emotionally unfit from induction prevented the wrecking of many lives and the fruitless ex- penditure of much time, effort, and money; and one can know surely that many lives were saved as the result of the one study alone which showed that the best night lookouts at sea were four times as proficient as the poorest. The Summary Technical Report of the Panel, prepared under the direction of the Panel Chief and authorized by him for publication, is a rec- ord of scientific accomplishment and of zealous effort by able men working to increase the ef- fectiveness of the nation’s military manpower in time of national peril. The members of the Panel have our gratitude. Vannevar Bush, Director Office of Scientific Research and Development J. B. Conant, Chairman National Defense Research Committee RESTRICTED FOREWORD WHAT ARE the major problems which the Army and Navy must solve if sound strategy and tactics are to win a war? One is the problem of personnel. The Navy’s Bureau of Naval Personnel and the General Staff evolved by years of military experience in the Army attest to the importance of personnel problems. The names of the General Staff divi- sions themselves indicate the extent of Service interest in the efficient use of manpower—G-l, Personnel; G-2, Intelligence; G-3, Training; and G-4, Supply. The interest in man is peren- nial. New methods of warfare, new explosives, new devices appear on the military scene, serve their period of usefulness, and become obsolete; the officer and the enlisted man remain. During the war years 1940 to 1942 the Serv- ices turned to scientists for help on personnel problems just as they did for help on materiel problems. Psychological sections were organ- ized in various branches of the Army and Navy. In 1940, the Adjutant General set up a Person- nel Procedures Section. In 1941, the Air Sur- geon created a Psychological Division. Early in 1942, the Navy’s Bureau of Medicine and Surgery established an Aviation Psychology Section. Later that year the Bureau of Naval Personnel organized a personnel research pro- gram. Within the National Defense Research Com- mittee [NDRC] civilian scientists and engi- neers, engaged in research and development of military devices, also turned to the psychologist. Man is such an important factor in the success- ful use of military equipment that there was early recognition of the need to study the man- machine combination, not the machine alone and not the man alone. Research on fire control, underwater sound, and communications equip- ment included psychological research on the human operators of these types of equipment. In night operations, too, the significance of the human factor was evident, and NDRC under- took psychological research on night vision. These scattered activities in and out of the Services proved the value of psychological re- search in meeting some Army and Navy prob- lems. Measurable increases in the efficiency of personnel followed the application of psycho- logical techniques. These improvements were of interest to nearly all branches of the Services and to a number of divisions of NDRC. The result was that psychological research units came to be sponsored by a remarkable variety of organizations: administrative, medical, engi- neering, physical, and mathematical. In spite of the diversity of its sponsorship, psychological research was always psycholog- ical, studying the man in relation to the ma- chines he had to operate. Experience proved that the same knowledge and the same experi- mental methods helped to solve military prob- lems which at first appeared quite unlike each other. Thus the merits of a single coordinating psychological research unit gradually became apparent. There was need for such a group to encourage the development of an overall, pro- fessional approach to military psychology. There was need for such a group, not to take over the many scattered groups already in ex- istence, but to supplement their work by enter- ing into fields which their limited directives and cognizances prevented them from entering. In direct recognition of this need the Navy Department, with the War Department concur- ring, on June 4, 1942 requested NDRC to estab- lish the Committee on Service Personnel— Selection and Training. This committee was formed in the National Research Council under an NDRC contract. The civilian members of the Committee consisted of John M. Stalnaker, Chairman, George K. Bennett, Leonard Car- michael, Clarence H. Graham, and Morris S. Viteles. Walter V. Bingham and P, E. McDowell served as Army and Navy representatives. The writer was the committee’s executive secretary. He was assisted by John L. Kennedy. In October 1943, the Committee on Service Personnel became the Applied Psychology Panel of NDRC. The Panel membership consisted of the civilian members of the predecessor com- mittee with the addition of Walter S. Hunter, who served as Chief of the Panel until the Japanese surrender, and Dael Wolfle. John L. RESTRICTED FOREWORD Kennedy and the writer were the technical aides. Later, Dael Wolfle also became a tech- nical aide. The first work of the Committee was almost entirely directed toward the solution of prob- lems which had high priority because of the shortage of manpower or the importance of the job itself, but which were not being studied by preexisting groups of psychologists. These spot programs concerned the selection, classification, and training of personnel. As the work pro- gressed, emphasis gradually shifted to attacks on more fundamental problems of simplifying selection, classification, and training procedures through improvements in the design and oper- ation of equipment. In the course of this shift many of the psychological projects came to be liaison as well as research centers, in which engineers, scientists, personnel men, and com- bat officers brought together their problems of man-machine efficiency and developed coordi- nated research programs on many aspects of military psychology. The experimental development and evalua- tion of personnel procedures led to marked improvement in Service efficiency. The two vol- umes of this Summary Technical Report contain the evidence that scientific methods can solve personnel problems in gunnery, fire control, communications, radar, and night operations; in surface, subsurface, ground, aerial, and am- phibious units; in elementary, basic, advanced, and operational training. The reader who studies the detailed record contained in the Panel’s Summary Technical Report will note one characteristic which is common to all successful research in military psychology. Studies of classification, training, and equipment, diverse though these subjects may appear to be, all rely on a common tool, the measurement of human performance on the job. The psychologist calls such a measurement the “criterion.” The criterion is the measuring stick with which he evaluates the success of a proposed test or procedure. Psychological tests, training aids and procedures, and designs for equipment, however sound they may be from a theoretical point of view, are valuable only in proportion as they improve human perform- ance. That performance must be measured if one is to know whether or not improvement has actually occurred. Three examples of the use of the criterion in military psychology are given in order to define the nature of psychological science today. When the Army and Navy asked the Panel to construct a test for the selection of radio code operators, the first problem was to develop a sensible, reliable, accurate, and objective cri- terion measure of men’s ability to receive code. Without such a criterion measure, there was no basis for knowing which of the already existing code aptitude tests was best or of knowing whether new tests were any better than the old ones. The development of a good criterion made it possible to construct a new code aptitude test, better than any previous ones. The new test was adopted by both Army and Navy. When the Navy asked the Panel to evaluate several procedures for transmitting code mes- sages over radar systems, the same kind of questions had to be answered. In this case, how- ever, the problem was to define the best rate at which to send messages and the best way of training radar operators to receive messages on their oscilloscopes. A solution based on fact rather than on opinion was possible only by measuring the proficiency of various groups of radar operators in receiving code at various rates after various kinds of training. When the Army Air Forces requested the Panel to improve the controls of the B-29 gun- sight, the question of a criterion arose again. An endless variety of controls was possible. Each suggestion had its advocates, but because the accuracy of aerial gunnery was so difficult to measure, choice among the alternatives was based on opinion instead of fact. The controls of a standard B-29 gunsight were attached to writing pens on a strip of paper. The paper moved at a constant speed under the pens. When the gunsight moved, the pens moved proportionally. With this device the gunner tracked a synthetic target. The ink rec- ord furnished a measure of the proficiency of the gunner. With this new criterion, one defective control was spotted at once. Each time that a gunner pressed his trigger the gunsight jerked 10 or RESTRICTED FOREWORD even 20 mils to one side or the other. In the remote control gunnery system of the B-29, the effect of pressing the trigger was not due to recoil of the guns. It was due to the relation between the gunner and his gun controls. The controls were improperly arranged for good triggering performance. The record of gunner performance permitted the evaluation of new controls. Choice of the best controls in terms of their relation to gunner performance was possible. The new record also permitted the Panel to recommend a new op- erating procedure for gunners using the stand- ard sight: Fire continuously when once you have begun to fire. It stimulated the develop- ment of an automatic burst-control trigger which fired the guns intermittently even though the gunner pressed the trigger continuously. The record permitted the evaluation of alterna- tive methods of training gunners. The same measure is available as a criterion for evaluat- ing aptitude tests for gunners. Throughout this Summary Technical Report the problem of securing an objective and re- liable measure of performance on the job recurs again and again. Whether the problem is to classify men, to train them, or to adjust operat- ing conditions for maximal efficiency, success depends upon the ability to measure accurately what men do. Considerable ingenuity was shown during World War II in developing measures of pro- ficiency, but too many of those measures were devised for school and training use. More effec- tive work could have been done if more oppor- tunities had been created for studying combat performance, the ultimate goal of all military research. In fact, the only phase of military operations which has so far not been directly studied on a large scale by the psychologist is combat itself. Better selection, more realistic training, and more efficient operating proce- dures will become available when combat per- formance is measured and used as a criterion. In the years to come the Army and Navy may have practically any kind of research program in the field of military psychology that they de- sire. Psychological work on military personnel during World War II varied from none at all to a thorough research analysis of the whole inter- related set of problems involved in selecting men for special duty, training those men for the duties they were to perform, and modifying the equipment in order to make it easier to operate and easier to learn to operate. The various pos- sibilities for the future are enumerated below. Prior to World War II it was the frequent ex- perience of psychologists interested in military problems to be told that the Army and Navy welcomed the opportunity to consult with them but needed no research assistance. This atti- tude persisted throughout the war years in the minds of some officers but was by no means universal, for research was not only welcomed but eagerly sought in a number of commands and on a number of problems. Both Army and Navy do know a great deal about personnel. Both have a long history of ex- perience and a record of success. In addition, both have the benefit of recent advice from ex- pert consultants and the results of wartime re- search studies to aid in handling personnel dur- ing peacetime. It might therefore be decided, either as a matter of general policy or for a par- ticular field, that no further research is neces- sary. ' The Applied Psychology Panel on a number of occasions was asked to furnish a man well acquainted with a particular type of military problem who could serve as an expert con- sultant for a few days’ time. During these short periods of consultation the man was expected to survey, for instance, a training installation, ob- serving the instruction and recommending im- provements in instructional material, course or- ganization, lesson plans, and examinations. The results of such consultations varied, de- pending upon the experience of the consultant and upon the similarity between the problems involved in the military situation and those with which he had had previous experience. One successful instance was the Applied Psy- chology Panel’s work on improving the training of Combat Information Center [CIC] personnel under Commander Operational Training Com- mand, United States Pacific Fleet [COTCPac]. In this case it was possible to provide COTCPac with a consultant who had wide and varied ex- perience with radar equipment, with radar op- erator training in the Services, and with educa- RESTRICTED X FOREWORD tional methods. In a few days’ time spent at each CIC school, this consultant was able to eliminate a number of difficulties and to suggest improvements which were eagerly received by the instructors in charge. The usefulness of consultation depends on the degree of familiarity of the advisor with actual Service conditions. The best of men may come into a Service situation with no direct experi- ence, or experience only from the remote past, of the details of Service operations. He can deal only in generalities. These do not compare in value with the specific, detailed recommenda- tions which are possible to an expert who has had the benefit of months of research work in a training or operating area. Consultation on special problems is the sec- ond type of opportunity available to the Army and Navy. As a third possibility, psychologists can be used to maintain already established personnel objectives and procedures. Psychological pro- cedures, like others, require continuing atten- tion or they gradually become ineffective. Fail- ure to maintain personnel methods results in a loss of the values originally attained. A program of maintenance is necessary. If, however, research is limited to maintenance alone, any mistake in original plans may persist unchallenged and new developments may pass unheeded. The final type of study which may be con- ducted consists of a thoroughgoing continuing research analysis of three interrelated sets of problems. These problems are the ones involved in the classification of personnel for particular duties, the problems involved in their training, and the problems involved in designing the equipment which they use in order to make both learning and operation easier and more effi- cient.'1 Working simultaneously on selection, train- ing, and equipment problems is the most effi- cient type of program and may normally be ex- pected to lead to the greatest improvements. As shown above, the research methods are the same for all three. And the three kinds of problems are all centered in the human operator; they are necessarily interrelated. Changing selection standards may permit a shorter, or require a longer, training period. Simplifying the opera- tion of a piece of equipment may allow a relaxa- tion of selection standards and require a modi- fication of training plans. Working simultane- ously on selection, training, and equipment al- lows for the efficient coordination of plans con- cerning all three problems. It seems clear that the most effective research for the future Army and Navy will be research on problems related to new equipment. Research on the classification and training of personnel and on methods of operation of equipment will be most effective if results are available when new devices roll off the production line. If studies of equipment design are completed be- fore production begins it will be unnecessary to suggest field modifications or second models. The best time for psychological research is be- fore the need for it is obvious. It is only neces- sary to add one or a few additional mock-ups and preproduction models to the five or six now commonly constructed to make psychological re- search possible in the preproduction period. The direction of psychological research to- ward new equipment suggests that psychologi- cal research should be associated with other re- search and development work. Only by associat- ing psychological research with all research will there be automatic provision that personnel re- quirements will be subjected to intensive study at the most effective time. The Army and the Navy can have any type of assistance they desire. If they desire a complete program they must secure men with adequate training in experimental psychology. Whether psychologists will accept Service positions de- pends upon the conditions offered for experi- mental work. Psychologists will not enter the Army or Navy during peacetime unless there is good as- surance that they will be able to work effectively on psychological problems. That assurance de- pends upon freedom from the danger of being a Since the Panel worked only on classification, train- ing, and equipment, no other aspects of military psy- chology are considered here. If a complete research pro- gram is contemplated, consideration should also be given to the inclusion of research on the abnormal individual and on the social psychology of morale, and psycholog- ical warfare. RESTRICTED FOREWORD sidetracked into routine administrative duty. It depends upon long-term planning; year-by- year commitments are acceptable during war- time, but not in planning peacetime research. It depends upon the ability of the military psy- chologist to maintain his status as a recognized member of the profession of psychology. It de- pends upon the creation of a single psychologi- cal command for psychological research within a Service or major branch of a Service. It de- pends upon a fairly high staff level of military supervision so that psychological planning and direction will be in the hands of officers with an appreciation of the broad implications of psy- chological research and with the authority to see that suitable research facilities are made available. Unless the Army and Navy are able to offer assurance that psychologists will be able to work effectively on psychological problems, the several hundred positions now being created for military psychologists will remain unfilled. At the time of writing this foreword only a handful of the thousand or more psychologists engaged in wartime psychological work have accepted permanent positions with the Services. Salaries, frequently 50 per cent higher than the universi- ties offer, are not attractive unless they are ac- companied by opportunities to work effectively and continuously on important psychological problems, opportunities to see the results in- corporated into Service practice, and opportuni- ties to maintain close relations with civilian colleagues. The Army and Navy can have any type of assistance they desire. Whether either Service will have an effective research program in mili- tary psychology depends upon the availability of personnel with adequate training in experi- mental psychology. Competent psychologists will be available when satisfactory conditions for experimental work are known to exist. Service interest in man is perennial. Service recognition of this fact is illustrated by the words of a battle-scarred and ribbon-covered Admiral appointed to demonstrate the mechani- cal and electronic marvels of the USS Missouri to a group of distinguished scientists from NDRC. His opening words, as he swept his hand over the deck of the pride of the fleet, were these: “Twenty-five hundred officers and men: gentlemen, twenty-five hundred sources of er- ror/’ It was the purpose of the Applied Psychology Panel to help the Army and Navy reduce human error during World War II. The success of fu- ture efforts to reduce human error—the success of the future development of military psychol- ogy—rests with the Army and the Navy. Charles W. Bray Chief, Applied Psychology Panel RESTRICTED PREFACE The Summary Technical Report is a syste- matic account of the work done under the direction of the Applied Psychology Panel. Vol- ume I describes selection and classification of military personnel; Volume II describes mili- tary training and the human factors involved in the design and operation of military equip- ment. In each of these three fields—selection and classification, training, and the design and operation of military equipment—the work ac- tually done and the effects of that work on mili- tary practice are described. In the foreword to this volume Dr. Charles W. Bray, Chief of the Applied Psychology Panel, has described the Panel’s recommendations for future research in military psychology. This volume and the Panel’s recommenda- tions for future work on classification are sum- marized in Chapter 1. Chapters 2 to 13 give de- tails of the Panel’s work on the selection and classification of military personnel. These chap- ters may be read in any order. Each gives an ac- count of one phase of the work. Chapters 14 and 15 will give the professional psychologist an un- derstanding of the psychological principles in- volved and the methods followed in the develop- ment and evaluation of personnel classification devices. In Volume II, Chapter 1 gives a brief factual summary of the Panel’s work on military train- ing and on the development of design improve- ments and standard operating procedures for several types of military equipment. Chapters 2 to 12 describe specific training studies. Chap- ters 13 to 17 cover general principles of military training. Chapters 18 to 23 discuss the design and operation of special types of military equip- ment. The two final chapters, 24 and 25, present some guiding principles for future equipment development. Cross references are given by means of sec- tion numbers, for example 11.3.5, in which the 11 refers to the chapter, 3 to the third major division of Chapter 11, and 5 to the fifth section of the third division of Chapter 11. Commonly- used abbreviations are explained in a glossary at the end of the volume. In writing this final summary a few tables have been recalculated from the original re- ports. In no case were the changes large enough to alter conclusions or recommendations. The Applied Psychology Panel has had help from many sources in preparing this final ac- count of its work. The Army, the Navy, the Applied Psychology Panel contractors, and the Bausch and Lomb Company have provided pho- tographs to illustrate many of the devices and procedures discussed. To each of these we ex- press our thanks. The author of each chapter is named in the table of contents and at the beginning of the chapter. Though all chapters are based upon the original reports prepared by the contrac- tors, many of them were finally written by the editor. Approximately half of the chapters were written by men who spent the war years in the field, working on problems their chapters sum- marize. The Panel and the editor are very appre- ciative of the time and effort devoted by these men to writing firsthand accounts of their work. Credit is not given in the Summary Technical Report to the individual psychologists who actu- ally made the contributions here reported. Their specific contributions can be discovered only by studying the original reports listed in the bibli- ographies. The Applied Psychology Panel ex- presses its sincere appreciation to these men for their individual contributions and for their ef- fective teamwork. These two volumes record their achievement. Dael Wolfle Editor RESTRICTED CONTENTS CHAPTER PAGE 1 Summary by Dael Wolfle 1 PART I SELECTION OF MILITARY SPECIALISTS 2 The United States Navy Basic Classification Test Battery by Dael Wolfie 10 3 The Selection of Officers by Dael Wolfie . . 27 4 Eliminating the Emotionally Unfit by Dael Wolfie 36 5 Determining Vocational Interests by Dael Wolfie 53 6 The Selection of Radio Code Operators by Dael Wolfie 58 7 Selecting Radar Operators by Donald B. Lindsley 63 8 Selecting Stereoscopic Rangefinder and Heightfinder Operators by William E. Kap- pauf, Jr 69 9 Selecting Night Lookouts by William E. Kap- panf, Jr 93 10 Selecting Navy Telephone Talkers by Dael Wolfie 100 PART II CLASSIFICATION OF MILITARY PERSONNEL 11 Principles and Devices for Military Classifica- tion by John L. Kennedy 103 12 Organization of a Classification Program for Recruits by Norman Frederiksen . . . . 118 13 Organization of an Advanced Classification Program by Norman Frederiksen . . . . 125 RESTRICTED XVI CONTENTS CHAPTER PAGE PART III PSYCHOLOGICAL PRINCIPLES IN THE DEVELOPMENT AND EVALUA TION OF CLASSIFICATION METHODS AND DEVICES 14 Construction and Standardization of Group Tests by Norman Erederiksen 134 15 Contributions to Test Theory and Test Ad- ministration by Dael Wolfie 147 Glossary 155 Bibliography 157 OSRD Appointees 165 Contract Numbers 166 Service Project Numbers 167 Appendix 171 Index 289 RESTRICTED Chapter 1 SUMMARY By Dael Wolfle 11 INTRODUCTION This chapter is a summary of the work of the Applied Psychology Panel, NDRC, on problems of military selection and classification —the work reported in the remaining chapters of this volume. The chapter is organized into four sections. The first outlines the Panel’s con- tributions to the development of tests for use in selecting military specialists. The second de- scribes the Navy classification program and the Panel’s contribution to it. The third describes the methods employed in the construction of tests for the selection and classification of mili- tary personnel. The fourth brings together the Panel’s recommendations for future develop- ment of military classification processes. The subtitles of the following sections include the numbers of the chapters summarized in each. 12 TESTS FOR SELECTING MILITARY PERSONNEL The United States Navy Basic Classification Test Battery (Chapter 2) The U. S. Navy Basic Classification Test Battery consists of five tests. Project N-106 of the Applied Psychology Panel and the Test and Research Section of the Bureau of Naval Personnel cooperated in their development. The Test and Research Section had major respon- sibility for developing two of the tests, the Mechanical Aptitude Test (MAT) and the Me- chanical Knowledge Test. The project assumed major responsibility for the initial development of the other three, the General Classification Test (GCT), the Reading Test (R), and the Arithmetical Reasoning Test (AR). The project developed two forms of each of these tests, pro- duced an additional form of the GCT suitable for administration in the fleet, and prepared an additional test named the Arithmetical Com- putation Test (AC) for possible use in future general selection batteries. Statistical information regarding the GCT, R, AR, and AC tests (and for all tests devel- oped by the Panel) is given in Table 1. The first three, as part of the Basic Classification Battery, were used from the time of their com- pletion at the end of 1942 until the end of World War II, in the initial classification and assign- ment of all Navy recruits. 1.2.2 The Officer Qualification Test (Chapter 3) The Bureau of Naval Personnel requested a test which could be used to select those officer applicants who had the ability to complete Navy officer training courses satisfactorily. The re- quest was met by developing three forms of the Officer Qualification Test (OQT). The OQT was used in the Offices of Naval Officer Procurement until officer recruitment ended at the close of World War II. 12'3 Predicting Combat Leadership (Chapter 3) In fulfillment of a request from the Office of the Adjutant General, War Department, the Applied Psychology Panel conducted a prelim- inary study of the prediction of combat leader- ship. Two classes of infantry officer candidates and two classes of field artillery officer candi- dates were studied. In addition, combat ratings of 176 infantry company officers were compared with several types of officer candidate school (OCS) records. Within the already-selected group of students in infantry OCS, it was found that none of the items of information obtainable before or upon entrance to the school was highly predic- RESTRICTED 1 2 SUMMARY Test No. of forms Purpose Reliability Validity Service use Details in chapter General Classification Test (GCT) 2 For use in general classifica- tion of naval recruits to decide on type of duty or special training .94 .32-.63* .27-.63 Routine Navy use (NavPers 16502) 2 Reading (R) 2 .85 .32-.59* .26-.46 Routine Navy use (NavPers 16512) 2 Arithmetical Reasoning (AR) 2 .83 .28-.63* .27-.51 Routine Navy use (NavPers 16512) 2 Arithmetical Computation (AC) 1 Experimental test for pos- sible inclusion in Basic Battery ,33-.69f Experimental use only; not adopted 2 Officer Qualification Test (OQT) 3 To predict ability to com- plete Navy officer training courses satisfactorily .90 .48-.511 .33-.47 Routine Navy use (NavPers 16561, 16563) 3 Personal Inventory (PI) 3 To identify emotionally un- stable men likely to be dis- charged for psychiatric cause .66-.92 See Chapter 4 Routine use by Navy (NavPers 16845), U.S. Coast Guard, U. S. Maritime Service. Adapted by AAF 4 Vocational Interest Inventory 1 To determine interests which might be used as an aid in military classification .77-.94§ Unde- termined None 5 Radio Code Aptitude Speed of Response (SOR) 1 To select men for training as radio code operators .50 Adopted for routine use by Army as A rmy Radio Code A ptitude Test, 1944, ARC-1, and by Navy as Radio Code Test: Speed of Response, Form 2. 6 Experimental forms of 13 tests designed to select radar operators .81-.98 Two of the 13 included in Navy CIC Aptitude Test, Form 2 (NavPers 16980) which had validity of .45-.46 7 Speech Interview Un- limited To rate ability of men for duty as shipboard telephone talkers Unde- termined Adopted by Navy for routine use 10 ♦There is, of course, no single validity for a test. The upper pair of values given are the highest and the lowest correlations found between scores on Form 1 tests and grades in 13 elementary Service schools. The lower pair are the highest and lowest correlations between scores on Form 2 tests and grades in 12 elementary Service schools. The correlations have been corrected for curtailment in range of talent. fRange of correlations with grades in 12 elementary Service schools. JUpper figures are correlations with grades i Reserve Midshipman’s School. in Naval Training School (Indoctrination); lower figures are correlations with grades made by women in Naval §Reliabilities of the five scores, each representing interest in a different area of behavior, varied from .77 to .94. Table 1. Selection tests developed by projects of the Applied Psychology Panel. RESTRICTED TESTS FOR SELECTING MILITARY PERSONNEL 3 five of later performance. Scores on the Army General Classification Test and leadership qual- ities as indicated by former civilian positions were the two most predictive variables. Background information available upon en- trance to field artillery OCS was generally more useful in predicting success in school than was true of infantry OCS. Previous education and grades on mathematics examinations showed useful relationships (correlations of .33 to .48) with success in field artillery OCS. There was fair agreement between combat ratings and leadership ratings assigned while in infantry OCS. Of the other variables studied, only age had predictive value for combat suc- cess. The superior and excellent officers on the whole came from the age group 22 to 28, Pro- portionately fewer men below 22 or above 28 were rated superior or excellent by their regi- mental commanders, executive officers, or bat- talion commanders. 1-2-4 Eliminating the Emotionally Unlit (Chapter 4) The Personal Inventory (PI) was developed in order to have available a device for making quick identification of those emotionally un- stable men who were likely to break down under the stress of hazardous duty. Three forms of the PI were developed. 1. A long form containing 145 items in each of which the man taking the test was forced to choose between two short descriptive state- ments, selecting the one which he felt more appropriately applied to him. 2. A short form consisting of 20 items se- lected from the long form. The 20 selected were those which experience had shown to be most useful in discriminating between acceptable men and men later discharged for psychiatric reasons. The correlation between the long and short forms was .84, 3. A long form suitable for use with officers. The PI was used as a preliminary screening device to identify the men most likely to be given subsequent psychiatric discharges. When the flow of men was too great to permit careful psychiatric interviews of all men, the PI identi- tied those who had to be interviewed most care- fully. Validity of the PI was determined by measuring the success with which it could select in advance those men whom the psychiatrists later decided unfit for duty. Its usefulness as a screen consisted of reducing the number of men who had to be interviewed in order to find a given proportion of the men who would later be discharged. The number to be interviewed could vary depending upon the amount of inter- view time available. For example, the PI could select a group of 16 per cent of the total and in that 16 per cent include 60 per cent of those given discharges. If time permitted interview- ing more men, the PI could select a group consisting of 42 per cent of the total and in the 42 per cent include over 75 per cent of those given discharges. Preliminary evidence indicated that the PI could contribute to improved selection of am- phibious forces officers, paratroopers, and avi- ation personnel. Its value in selecting Marine Corps officer candidates was slight. The short form of the PI was used by the Navy as a preliminary psychiatric screen. The long form was used by the Coast Guard and the U. S. Maritime Commission. The Army Air Forces developed a special adaptation for their use. 1-2-0 Determining Vocational Interests (Chapter 5) At the request of the Adjutant General, War Department, an inventory of interests was con- structed for possible use in the classification of military personnel. The inventory was intended to secure information on a wide variety of interests, for example, vocational, social, recre- ational, scientific, and religious interests. In- formation of this type could provide a valuable supplement to that obtained by interviews, vo- cational history, or standardized tests of ability in classifying men for military duty. The interest inventory, a scoring key, and detailed statistical information were turned over to the Adjutant General’s Office at their request, for determination of its actual useful- ness in military situations. No report has been received from that office. RESTRICTED 4 SUMMARY Selecting Radio Code Operators (Chapter 6) At Navy request the Applied Psychology Panel developed an improved test of radio code aptitude. The test first required the subject to learn three simple characters of International Morse Code and then measured his ability to differentiate those characters as they were sent at faster and faster speeds. Instructions, the learning part of the test, and the test itself were all recorded on phonograph records to en- sure standardized administration. The test was officially adopted by both Army and Navy for use in the selection of men for radio code training. Selecting Radar Operators (Chapter 7) A number of tests were designed for select- ing and screening radar operators. The tests were intended to measure aptitude for the visual tasks of reading and interpreting oscil- loscope patterns. Two of the tests were adopted after revision by the Navy for use as two of the three parts of the Combat Information Center Aptitude Test, Form 2 (NavPers 16980). The third part was developed by the University of California Division of War Re- search under NDRC Section 6.1. The total test showed correlations of .45, .55, and .56 with final grades in three successive classes at Naval training school (tactical radar). 1-2'8 Selecting Stereoscopic Rangefinder and Heightfinder Operators (Chapter 8) Standards for the selection of stereoscopic heightfinder operators were developed for, rec- ommended to, and adopted by the Army. A Panel project demonstrated the validity of these standards. Assistance was given to the Army in setting up stereoscopic testing centers where men to be trained as heightfinder operators were selected. When the Navy established a school for train- ing men for the rate of fire controlmen (R), the Army heightfinder standards were used as the basis for setting selection requirements for rangefinder operators. The Applied Psychology Panel work was extended at that time to in- clude studies of the selection and training of rangefinder operators. Improved selection standards were recommended to the Navy and adopted for general use. The Project staff trained Navy personnel for duty in the two testing centers where prospective rangefinder operators were selected. In the course of developing selection stand- ards for stereoscopic observers, information was obtained on the reliability, validity, and intercorrelations of a number of visual tests. Intercorrelations among the tests purportedly measuring stereoscopic ability were shown to be conspicuously lower than the test reliabili- ties. This finding suggests that the tests do not measure exactly the same visual function and that each test must measure at least some ele- ments which are specific to itself alone. A thor- ough study of military visual requirements and of improved methods of visual testing was rec- ommended. Selecting Night Lookouts (Chapter 9) Studies on the prediction of night lookout performance showed that the tests of night vision being used had low reliability and very low validity. One study, for example, showed no relation between scores on the radium plaque adaptometer and measured performance of the men as night lookouts. The Bureau of Naval Personnel directed thereafter that scores on the radium plaque adaptometer should not be recorded on the Enlisted Personnel Qualifica- tions Card. 1-210 Selecting Navy Telephone Talkers (Chapter 10) A speech interview to rate the potential abil- ity of men as shipboard telephone talkers was developed. The interview was conducted over Navy sound-powered phones. It included the pronunciation of numbers and of all common RESTRICTED THE CLASSIFICATION OF MILITARY PERSONNEL 5 American speech sounds, repetition of com- mands, and extemporaneous description. A training course was developed to teach classification petty officers to administer the speech interview and to rate men as “well qualified,” “qualified,” or “not qualified” for duty as telephone talkers. The speech interview was adopted by the Bureau of Naval Personnel for routine use in the classification of enlisted personnel. Summary of Tests Developed under the Applied Psychology Panel Table 1 is a summary of the tests developed by projects of the Applied Psychology Panel for use in military selection. It gives pertinent statistical information and indicates the use made by the Army or Navy of each test. 13 THE CLASSIFICATION OF MILITARY PERSONNEL The tests described in Chapters 2 to 10 were developed to aid in the selection of men who would be successful in particular military spe- cialties or to aid in eliminating those unfit for special types of military duty. If the Army or Navy had an unlimited pool of men from which to select its specialists, each of these tests could select a group of men most of whom would sat- isfactorily learn their new duties. With a lim- ited manpower supply, the military task is more complicated than that: all available men must be allocated. This process requires using every man and attempting to assign each to the type of duty in which he will be most successful. The Objectives of Classification (Chapter 11) The objectives of classification are to select able men for special training, to assign each man to the job he can do best, and to divide available manpower equitably among the vari- ous specialties and among the various units of a military force. The Applied Psychology Panel cooperated with the Bureau of Naval Personnel in three programs designed to develop pro- cedures for achieving these objectives. First was the classification of the crew of USS New Jersey. Following this there were larger projects to classify men for the Amphibious Training Command of the U. S. Atlantic Fleet and for the crews of destroyers being manned on the West Coast. On the basis of the success of the three programs, the Commander-in- Chief, United States Fleet, directed the Chief of the Bureau of Naval Personnel to develop similar classification programs for fleet-wide application. 1.3.2 Devices to Aid Military Classifica- tion (Chapter 11) Four types of devices were developed to aid in the process of military classification. One was the series of tests described above. The second was a series of special devices to aid the interviewer in securing relevant informa- tion about a man’s experience and ability. The third was a mechanical system used on ship- board for sorting personnel record cards in order to locate quickly the men who possessed needed skills or traits. The fourth was a device, called the selectometer, which automatically weights and combines a man’s scores (or other measures) on each of the factors considered important in selecting men for a particular job. The device differs from a human classification interviewer in being more consistent in the importance it attaches to each variable, in not forgetting any of the important traits, and in being able to weight and combine scores used in predicting a man’s potential success in a number of different jobs simultaneously. Sev- eral models of the selectometer were built for use in Navy classification centers. 1.3.3 Organization of a Classification Program for Recruits (Chapter 12) The procedures followed during World War II in classifying and assigning Navy recruits are briefly described. Five improvements in the classification procedures are recommended. 1. Using a primary test battery to separate the men of school quality from those who should be assigned to general detail and a secondary battery to determine the particular school to RESTRICTED 6 SUMMARY which each man in the first group should be assigned. 2. Improving the interview by standardizing the weighting assigned to various factors on which the interviewer’s recommendation should be based. 3. Improving the basis for filling school quotas by more general use of assignment pools and by better determination of proper priorities in filling quotas. 4. Continuous research to keep classification tests and methods abreast of changing military requirements. 5. Indoctrinating regular officers in modern methods of personnel classification. Organization of an Advanced Classification Program (Chapter 13) The methods used by the Navy in advanced classification and reclassification of enlisted personnel are briefly described. Five recom- mendations for improving advanced classifica- tion are offered. 1. More extensive construction and use of objective methods of determining the actual proficiency with which a man can perform the duties of a particular billet. 2. Improvement in the methods of rating shipboard proficiency. 3. Use of brief oral tests, called work readi- ness tests, to determine how much a man knows about the details of a particular type of duty. 4. The indoctrination of regular officers in the methods and information available for use in making shipboard assignments. 5. Improved methods for coding and filing personnel data to make those data more easily available and more likely to be used in ship and shore establishments. 14 METHODS OF TEST CONSTRUCTION AND VALIDATION 141 Construction and Standardization of Group Tests (Chapter 14) Test construction and validation are ana- lyzed in detailed outline. The first step consists of discovering what critical abilities or skills are required by the job for which the test is being constructed. It is further necessary to acquire at least a general notion of the charac- teristics of the group to whom the test will be given. Also, the conditions and limitations of the classifications program as a whole neces- sarily affect the new test and must be consid- ered in planning its details. After this background work is completed, the test itself can be written in experimental form. Detailed statistical analysis of the re- sponses made to each item by a sample group of men provides the information necessary to select items for the final test in terms of their difficulty and validity, to decide on time limits, to make certain that instructions are clear, and to arrange the items in proper order. 1-4 2 Contributions to Test Administra- tion and Test Theory (Chapter 15) An investigation of the accuracy with which aptitude tests were being scored in 1942 in the Navy revealed frequent and large errors. Meth- ods of controlling and checking scoring pro- cedures were devised. Subsequent investiga- tions showed that errors were greatly reduced in frequency and in size. Several innovations in test construction practice were tried out. They included: 1. A simple method of measuring the pre- mium which a test places on speed by compar- ing the number of testees who attempt the last item of the test with the number attempting the first one. 2. A method of estimating the validity of a test for a school population without waiting for training and performance measures to accumu- late after giving the test. Instead of waiting, the test is given to a graduating class and their test scores are correlated with final grades. Simultaneous administration of the test to a sample of entering students provides a check on the accuracy of the predicted validity coeffi- cient by permitting comparison of regression weights for entering and graduating classes. 3. A method for increasing the validity of a test by selecting items in terms of their corre- RESTRICTED THE FUTURE OF RESEARCH IN MILITARY CLASSIFICATION 7 lation with an external criterion such as suc- cess at the school, instead of their correlation with the total-test score. 15 THE FUTURE OF RESEARCH IN MILITARY CLASSIFICATION What remains to be done in the field of apti- tude testing for military use? The experience of the Applied Psychology Panel suggests three principal needs. 1. New tests are necessary. Tests now avail- able are not uniformly satisfactory. Better measures of certain types of ability need to be developed. In addition, changing equipment and changing methods increase or decrease the importance of careful selection of different kinds of specialists. Selection programs must keep up with these changes. 2. Psychological tests by their very nature require follow-up and review. They need main- tenance and recalibration just as a physical measuring device requires maintenance and recalibration. 3. The actual usefulness of tests during World War II was considerably less than their potential usefulness. A number of administra- tive problems must be solved before the Army or the Navy can derive full benefit from their research programs on aptitude measurement and personnel classification. These points are discussed below. New Tests Are Necessary Among the more general capacities for which better tests are needed are emotional fit- ness for combat, vocational interests, and visual and auditory capacity. Tests exist in each of these fields, but they are less effective than the general classification or mechanical aptitude tests. The special abilities for which better tests are needed will vary from time to time depend- ing upon the types of equipment to be used and the methods current for their use. For example, there is now reason to believe that the skill of the operator of a fire control director is rela- lively unimportant compared with the skill of the maintenance man responsible for calibra- tion and alignment of the parts of the antiair- craft battery. In general, as equipment is more highly developed, it becomes possible to sim- plify some aspects of the operator’s task. This simplification is frequently made at the ex- pense of complicating the maintenance and calibration difficulties of the equipment. With such a change, a larger portion of the military population can serve satisfactorily as oper- ators; their selection is therefore less critical. But the selection of maintenance personnel be- comes more critical. Because equipment changes, personnel re- quirements change. Because personnel require- ments change, some tests become obsolete and others become necessary. New tests will be necessary as long as new equipment and new methods of warfare continue to be developed. In developing new tests, more attention than was paid in World War II should be given to developing tests which are statistically inde- pendent of each other. The usefulness of a test in military classification depends, among other things, on its correlation with other tests also being used for classification purposes. Tests which have low correlations with other tests in a battery are more valuable than tests which have high correlations. Enough apt men for all specialties can be found only by having a test battery which picks out the special aptitudes of each man and furnishes a number of scores each of which is relatively unaffected by gen- eral ability. In developing new tests, more emphasis than was given in World War II should be placed on field or combat skill as a criterion for validat- ing the tests. Even at best, school grades are an unsatisfactory substitute for actual perform- ance records in the validation of selection tests. School grades were frequently used in World War II because they were easily available and because they were more reliable than combat records. However, the ultimate goal of both selection and training is the achievement of highly developed combat skill. Improving the ability to predict that skill requires continuous effort to secure more reliable combat perform- ance criteria. RESTRICTED 8 SUMMARY viewer had the test scores. He also had a good deal of additional information which he himself had collected during the interview. He should have been able to make better predictions using all this information than was possible with the scores on a single test. Evidently, most inter- viewers were too much impressed with the in- formation they had secured and gave it greater weight than they assigned to the relatively more objective and reliable evidence of the test scores. Better training of interviewers, the use of more objective and reliable methods of secur- ing information obtained in the interview, and standardized weighting of the variables to be considered in selecting an appropriate assign- ment will make it possible for the interviewer’s judgment to add to, rather than subtract from, the value of assignment by test scores alone. Methods for bringing about this improvement are described in Chapter 11. Personnel Policies Psychological tests are relatively new. Their military value, however, was demonstrated in World War I. They were widely used in World War II. Nevertheless, the administration of personnel policies, many of which grew up long ago, frequently rendered the test informa- tion useless. Two illustrations may be given. The first illustration arises from the opera- tion of the quota system in making assign- ments. Pools of men of all kinds are maintained at various Army and Navy centers. As new ships are commissioned, as new units are acti- vated, or as new classes enter a training school, men are drawn from these pools. Frequently the pool is not tapped until a few days before the quota must be filled. It is then necessary to take the best men available at that time and in that pool. This sometimes means that men not qualified for school training must be in- cluded in order to fill the quota. Sometimes it means that a man highly trained in one spe- cialty must be assigned to a totally new kind of duty because the pool is not large enough at the particular moment to fill the quota appropri- ately. As a specific example, the Army assigned a number of college graduates who had had special training for radar maintenance to duty 15 2 Aptitude Tests Require Follow-up and Review A test, like any precision instrument, must be adjusted and maintained as time passes. It cannot ordinarily be constructed, put to use, and allowed to continue in use without regular review and adjustment. An item which is satis- factory in difficulty when most recruits are coming from one section of the population may be too easy or too hard when the source of re- cruits changes. A change in the educational system of the country, for example an increase in the number of high school graduates, may make a formerly satisfactory test too easy. Susceptibility to these influences can be mini- mized, but it cannot be avoided completely. Only continuous watch over the tests and peri- odic statistical analysis of the distributions of item difficulties, test scores, and validity coeffi- cients can keep them at maximal usefulness. 1‘5'3 Administrative Use of Tests During World War II several factors de- creased the usefulness of the information ob- tained by aptitude tests. Two deserve special mention: the fallibility of interviewers’ judg- ment, and ignorance regarding the usefulness and availability of test scores. Interviewers’ Judgment Evidence was collected at the end of World War II which indicated that the work of the classification interviewer could be modified with resulting increase in the value of his recommendations. The evidence came from a direct comparison of the accuracy of interview- ers in predicting training school performance with the accuracy of prediction based on test scores alone.1 The test scores alone were clearly superior. The correlation between scores on a single test, the Mechanical Knowledge Test, and final grades of 3,496 men in electrician’s mate schools was .50; the correlation between interviewers’ recommendations and final grades of the same men in the same schools was .41. In making his recommendations the inter- RESTRICTED THE FUTURE OF RESEARCH IN MILITARY CLASSIFICATION 9 as truck drivers and as mess sergeants because there were more radar men and fewer truck drivers and mess sergeants than needed at that particular place and time. The dates or approximate dates of commis- sioning a new ship, activating a new unit, or starting a new class are usually known for at least some weeks in advance. It is not easy to work out administrative procedures which will take advantage of this advance knowledge to minimize the number of misassignments re- sulting from the quota system. But advantage has occasionally been taken of it, and service efficiency was improved as a result. A second illustration of failure to utilize carefully collected information about each man’s abilities comes from an old shipboard practice. Men are frequently assigned to their bunks alphabetically, and that assignment is naturally one of the first things done for a man when he comes aboard ship. On shipboard men must have their bunks near their battle stations in order to meet surprise attack promptly. The result is that the position of a man’s name in the alphabet may have more to do with deter- mining his battle station than does all the in- formation available about his abilities. The total effect of administrative malprac- tice is unknown. The following quotation from an Army report, however, indicates that it is sometimes very large. “About March 1, 1945, the Armored Replacement Training Center completed training of its 150,000th replace- ment. Where these men have gone is one of the mysteries of the war. Certainly they have not gone to the units for which they were trained. Some units state that they have received no men with armored training; others comment upon the few that arrive.”2 To have most of 150,000 men, all specially trained for one type of duty, disappear into a variety of other assignments is evidence that there is still much to be done in the way of administration of sound personnel policies before maximal effect can be obtained from present personnel selec- tion and training programs. Better personnel procedures and better standardization of them will help overcome these difficulties. So will more stringent quality control policies. The greatest gain will be achieved when regular officers become as aware of the importance of getting the right man for a job as they are of getting the right ammunition for a gun. RESTRICTED Chapter 2 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY By Dael Wolflea Summary should be developed to aid in the tremendously increased assignment problems of a wartime Navy, A new and very small section was estab- lished in the Training Division, Bureau of Naval Personnel, for this purpose. At about the same time, the National Research Council’s Committee on Service Personnel—Selection and Training (which later became the Applied Psychology Panel) was requested to make a survey of the procedures used in the selection, classification, and training of Navy personnel and to submit its findings and recommenda- tions to the Bureau of Naval Personnel. The Bureau responded to this report1 by an imme- diate request to OSRD to establish a project on “Research and Development of the Navy’s Aptitude Testing Program.” In response to this request, Project N-106 was formed under a contract with the College Entrance Examina- tion Board, Princeton, New Jersey. 211 The Basic Classification Test Battery Through the joint and always closely co- operative work of the rapidly expanded Test and Research Section of the Bureau of Naval Personnel and Applied Psychology Panel Proj- ect N-106, there was constructed a battery of aptitude tests designed to improve efficiency in the placement of Navy recruits and to assist in the maximum utilization of the Navy’s man- power. These tests were collectively called the U. S. Navy Basic Classification Test Battery. The battery contains five tests printed in four test booklets. Those developed by the Applied Psychology Panel project are marked with an asterisk. *1. General Classification Test (GCT). This is a composite test of 100 items, divided into three subtests as follows: a. Sentence Completion subtest (30 items, 10 minutes). WHEN EXPANSION of the U. S. Navy began, it became apparent that the tests and pro- cedures being used to classify and assign en- listed personnel were out of date and inefficient. A Test and Research Section was established in the Bureau of Naval Personnel to improve the classification tests. At about the same time, the Applied Psychology Panel of NDRC was asked to establish a project for “Research and De- velopment of the Navy’s Aptitude Testing Pro- gram.” The Test and Research Section and the NDRC project, working in close cooperation, developed the U. S. Navy Basic Classification Test Battery. After it proved successful, two additional forms of the battery were prepared for use ashore. A fleet edition of the General Classification Test was also developed. Statistical analysis of the results of the new tests, in comparison with similar analyses of the tests in use before 1942, showed the new tests to be reliably superior to the old ones. Comparisons between test scores and grades made in a variety of Service schools showed that the tests predicted well which men would and which would not do well in different Serv- ice schools. 21 INTRODUCTION For a number of years preceding the war all Navy recruits were given a set of paper-and- pencil tests which were intended to help in the selection or rejection of applicants and in the assignment of successful volunteers. This set of tests remained unchanged for many years. In 1942 the Navy decided that its test program should be brought up-to-date and that tests a This chapter is based primarily on the work of the staff of NDRC Project N-106. 10 RESTRICTED THE OLD NAVY TEST BATTERY b. Opposites subtest (30 items, 8 min- utes) . c. Analogies subtest (40 items, 15 min- utes) . *2. Reading Test (30 items, 25 minutes). *3, Arithmetical Reasoning Test (AR) (30 items, 30 minutes). (Both the Reading and Arithmetical Reasoning Tests are bound in the same booklet.) 4. Mechanical Aptitude Test (MAT). This is a composite test of 129 items, divided into three subtests as follows: a. Block Counting subtest (45 items, 6 minutes). b. Mechanical Comprehension subtest (44 items, 20 minutes). c. Surface Development subtest (40 items, 8 minutes). 5. Mechanical Knowledge Test (MKT). This is a composite test of 135 items, of which 60 relate to electrical knowledge and 75 to me- chanical. Both an electrical and a mechanical score are obtained from the test. The total time allowed is 37 minutes. Responsibility for the construction of some of these tests was assigned to the project, while others were developed by the Bureau. Actually there was cooperative work on all tests. De- tailed discussions and planning by both groups preceded the development of all the tests. Tests constructed by either group were criticized by the other. In many respects the Applied Psy- chology Panel’s project on “Research and De- velopment of the Navy Aptitude Testing Pro- gram” served as a closely integrated part of the Test and Research Section of the Bureau of Naval Personnel. 22 THE OLD NAVY TEST BATTERY 2,2-1 Statistical Analysis of the Old Battery2’4’7 8 The first step in the construction of the Basic Classification Test Battery was a thorough study and statistical analysis of the tests for- merly used for recruit selection and assign- ment. These tests were: 1. The O’Rourke General Classification Test, Junior Grade, Form C and D, Navy Edition. 2. The O’Rourke Mechanical Aptitude Test. 3. U. S. Navy Standard Recruit Test in Arithmetic. 4. U. S. Navy Standard Recruit Test in English. 5. U. S. Navy Standard Recruit Test in Spelling. 6. U. S. Navy Radio Code Aptitude Test. The Bureau of Naval Personnel supplied the Applied Psychology Panel with answer sheets for these six tests from 43,539 recruits. The tests were given at all naval training stations which existed at that time. These stations were located at Great Lakes, Illinois; Newport, Rhode Island; San Diego, California; and Nor- folk, Virginia, All the men had been tested be- tween January and April 1942. A sample of these 43,539 papers was care- fully rescored by the project staff and then sub- mitted to detailed statistical analysis to de- termine the difficulty, reliability, and internal characteristics of the tests. Correlations with Service school grades were also computed to Table 1. Reliability coefficients of old Navy aptitude tests. Training Station GCT (Form B) GCT (Form C) Code Mechanical Aptitude Test Arithmetic English Spelling Newport .95 .95 * .94 .89 .91 .94 Norfolk .95 * .69 .94 .88 .89 .93 Great Lakes ♦ .94 .75 .94 .87 .86 .92 San Diego ♦ .96 .78 .95 .88 .92 .92 Number of items 100 100 78 162 20 100 50 Time allowed (in min) 60 60 55 25 30 20 ♦Information not available. RESTRICTED 12 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY determine the validity of the tests. Samples of the papers from different naval training sta- tions were compared for the information such comparisons could give on the Navy population. This work was all done as a background for the development of the U, S. Navy Basic Classi- fication Test Battery and to help in the neces- sary decisions as to the length, difficulty, type of item, etc., which should characterize the tests in the new battery. Difficulty4 The frequency distributions of scores indi- cated that the General Classification, English, and Spelling Tests were too easy to provide the most efficient differentiation. The Arithmetic and Mechanical Aptitude Tests were of suit- able average difficulty but failed to exhibit as great a concentration of scores in the middle range as would be desirable. With respect to item difficulty, the ideal form for the distribution of the difficulty of the items of a general purpose test is not known exactly. It seemed, however, that for tests whose items have a comparatively high correlation with the total test score, a distribution of item diffi- culties should be flatter than for tests with low item-test correlation. This relation was not found to hold in the tests under investigation. Reliability2 The odd-even reliability, corrected by the Spearman-Brown formula, of the individual tests making up the old Navy Battery was satisfactory for all tests except the test of code aptitude. The reliability coefficients for each test as determined at the four training stations are presented in Table 1. The two parts of the Mechanical Aptitude Test were each reliable, but the correlation be- tween them was only about .65 to .70, which made questionable the desirability of combin- ing the two parts into a single test score. INTERCOKREL ATIO NS4 The intercorrelations of five of the tests, omitting the Radio Code Aptitude Test, are given in Table 2. Some of the intercorrelations were high enough to suggest the desirability of omitting one or two of the tests. Item Analysis7 The detailed item analysis of the tests gave information on the usefulness of the individual items. Of the 11 different types of items contained in the General Classification Test, the five types with the highest average biserial r’s were, in descending order, proverbs, completion, oppo- sites, synonyms, and analogies. This finding Table 2. Intercorrelations of old Navy aptitude tests (iV = 500 for each station). GCT MAT Arith- metic English Spell- ing Aver- age GCT Newport .46 .51 .58 .58 .53 Norfolk .50 .61 .64 .60 .59 Great Lakes .42 .49 .54 .53 .50 San Diego .63 .56 .69 .66 .64 MAT Newport .29 .29 .22 .32 Norfolk .39 .32 .30 .38 Great Lakes .27 .29 .22 .30 San Diego .47 .53 .39 .51 Arithmetic Newport .46 .44 .43 Norfolk .49 .49 .50 Great Lakes .48 .52 .44 San Diego .57 .55 .54 English Newport Norfolk Great Lakes San Diego .54 .55 .56 .59 .47 .50 .47 .60 Spelling Newport Norfolk Great Lakes San Diego .45 .49 .46 .55 supported the view that items which definitely require some reflection or cogitation are, in general, superior to those which can be an- swered with little or no thinking. A careful study was made of items which showed exceptionally high item-test correla- tions at each station and those which showed exceptionally low item-test correlations. The detailed earmarks of good versus poor items were found not to be uniform from one test to another. It appeared that the reasons for a low biserial correlation with total score were gen- RESTRICTED THE OLD NAVY TEST BATTERY 13 erally more discernible than the reasons for a high correlation. For practical work in test construction this finding suggests that the very poor items of a test may often be caught and rejected simply by careful inspection without the need for an actual tryout. The opposite, however, was not true. No consistent charac- teristics of exceptionally good items could be found. Probably the most important general conclusion from this part of the study is that no prediction, however sagacious, can substi- tute adequately for the detailed quantitative information about an item that is yielded by an actual trial. The actual trial of items is espe- cially useful for detecting excessively close or excessively remote (nonfunctional) distractors and for the identification of the most effective items in a test. Validity8 In order to determine the validity of the tests comprising the old battery, the scores on each of the six tests were correlated with grades in six Service schools at the Great Lakes Naval Training Station. The Service schools were electrician’s mate, fire controlman, gunners mate, torpedoman, quartermaster, and signal- man. The principal results of the validity study were: 1. The Service school grades used as criteria were highly reliable. The reliability coefficients varied from .93 to .97. These coefficients were determined by correlating the sum of the grades for odd-numbered weeks with the sum of the grades for even-numbered weeks and then applying the Spearman-Brown correction formula. 2. The validity of the individual tests was generally low. Only five of the 86 validity co- efficients were as high as .29. These five were .54, .38, .37, .37, and .29. 3. The General Classification Test was a better predictor of school grades than any of the more specialized tests. Table 3. Summary of data on validity of old Navy aptitude tests. Group I Group II EM EC GM TM QM SM Tests Corr. Tests Corr. Tests Corr. Tests Corr. Tests Corr. Tests Corr. Best single GCT .54 Arith. .38 Arith. .27 MAT 24 Arith. .27 GCT .31 test (r) Spell. .27 Best single GCT .72 GCT .65 Arith. .42 MAT 28 Arith. .40 GCT .51 test (rc) Best single GCT .54 Arith. .38 Age .28 MAT 24 Arith. .27 GCT .31 variable (r) Spell. .27 Best combi- GCT-MAT .55 Arith.-GCT .32 RCA-Arith .34 nation of GCT-Arith. .55 GCT-Arith .44 Arith.-MAT.32 MAT-RCA 26 RCA-Spell. .34 GCT-RCA .36 two tests GCT-Eng. .55 (R) Best combi- GCT-MAT .55 nation of GCT-Arith. .55 Arith.-Age .46 GCT-Age .37 MAT-Age 31 Arith.-Age .36 GCT-RCA .36 two vari- GCT-Eng. .55 ables (R) GCT-Age .55 Best combi- GCT nation of Eng. .56 three tests MAT GCT GCT GCT Arith. GCT (R) Arith. .46 Arith. .34 MAT 27 Spell. .38 Spell. .41 GCT MAT MAT RCA RCA RCA Eng. .56 Arith Best combi- GCT GCT GCT GCT Arith. GCT nation of Eng. .57 Arith. .51 Arith. .40 MAT 31 Spell. .39 Spell. .41 three vari- Age Age Age Age Age RCA ables (R) RESTRICTED 14 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY 4. Best prediction was found in those schools which contained a large portion of academic work. 5. The best combination of two tests gave multiple correlations which on the average were only .04 higher than the best single test. The chief statistical results of the validity study are summarized in Table 3. The values of rc in the second line of the body of this table are the correlations corrected for curtailment of the distribution of the sample studied. The tests included in the new Basic Aptitude Test Battery and in supplementary Navy tests contain items of every type which this validity study suggested might be of value. tained. On all tests the highest average scores were made by the recruits at the Great Lakes Naval Training Station. The lowest average scores were all made by recruits at the Norfolk Naval Training Station, The data are pre- sented in Table 4. Comparisons based on a much smaller group of papers, which had been very carefully re- scored by the project, showed smaller differ- ences between the four naval training stations but did not alter the rank order. The number of cases involved was from 200 to 300 for each test from each station. The smaller population was selected at random from all the papers available from each station. In this smaller and more carefully scored sample the average score at Norfolk was generally from one third to one half a standard deviation below the Great Lakes average. Discrepancies in the cards on which the scores were recorded and gross differences be- tween the reports coming from the different stations suggested that differences in examina- tion and scoring procedures were large and made it appear probable that testing conditions had sometimes departed seriously from recom- mended procedures. 2 2 2 Revision of Mechanical Aptitude Test Upon request of the Standards and Curricu- lum Division, Bureau of Naval Personnel, de- tailed suggestions were made for the revision of the mechanical aptitude test then in use in the Navy.5 These suggestions were accepted and incorporated into new forms of a mechani- cal aptitude test which the Standards and Cur- riculum Division constructed as part of the Basic Battery. 23 EXPERIMENTAL TESTS In order to secure additional information on the types of tests which should be included in the basic classification test battery, an experi- mental battery of 11 tests was administered to entering classes in Gunners Mate, Fire Control- Table 4. Means and standard deviations of Navy apti- tude test scores for four stations (based on 43,000 cards). Test by stations Number of cases Mean Standard deviation GCT Newport 4,858 69.35 14.53 Norfolk 11,740 65.72 20.59 Great Lakes 15,878 77.02 13.16 San Diego 8,489 66.85 17.39 MAT Newport 4,837 46.25 15.31 Norfolk 10,798 33.83 25.86 Great Lakes 15,650 55.04 27.27 San Diego 7,866 51.54 16.82 Arithmetic Newport 4,864 48.17 23.89 Norfolk 11,753 44.75 26.80 Great Lakes 15,820 65.03 23.45 San Diego 7,808 47.83 25.00 English Newport 4,852 65.81 12.63 Norfolk 11,752 33.35 21.56 Great Lakes 15,831 71.32 11.65 San Diego 7,092 63.86 15.23 Spelling Newport 4,832 57.13 20.57 Norfolk 11,739 51.91 24.58 Great Lakes 15,199 67.71 19.26 San Diego 7,068 57.82 21.23 Variability in Navy Samples2- 4 Consistent differences were found in the average scores and in the distributions of scores reported from the four naval training stations from which answer sheets were ob- RESTRICTED CONSTRUCTION OF TESTS FOR THE NEW BATTERY 15 man, and Rangefinder Operator’s Schools at the Navy Yard, Washington, D. C.10 The 11 tests included a verbal test, tests of quantita- tive ability, spatial relations tests, and a test of mechanical aptitude. After the men had gradu- ated, their Service school grades were used as criteria in evaluating the 11 tests. As judged by predictive efficiency, the seven best tests turned out to be tests of arithmetical reasoning, arithmetical computation, vocabu- lary, mechanical comprehension, surface devel- opment, the old general classification test, and the old mechanical aptitude test. The new Navy Basic Classification Test Battery was designed to contain items similar to those in all these tests except arithmetical computation (see Sec- tion 2.4). A new test of arithmetical computa- tion was prepared later by the project (see Section 2.5), 21 CONSTRUCTION OF TESTS FOR THE NEW BATTERY A new General Classification Test, a new Reading Test, and a new test of Arithmetical Reasoning were constructed by Applied Psychol- ogy Panel Project N-1063 for inclusion in the new Basic Classification Test Battery. Certain specifications for these new tests were stated by the Bureau of Naval Personnel. These were: The general classification test was to include only items involving verbal ability. Types of items suggested were sentence completion, opposites, and analogies. A sentence completion item is a sentence with one word omitted; the task is to choose the one of several words presented which best fits in with all parts of the sentence. In such items, it should be necessary for the candidate to sense the implications of all parts of the sentence in order to choose the correct word. The choice should not depend simply on vo- cabulary knowledge, nor should the test meas- ure good word usage in a purely literary sense. An opposites item consists of a key word followed by several other words, one of which is the opposite of the key word. The task is simply to choose the word which means the opposite of the key word. Highly technical words were to be avoided. Words likely to occur in Navy manuals were recommended for use. An analogies item is a statement of an an- alogy with the last word missing. The task is to choose, from a list of several words, the one which best completes the analogy. Items of this type should depend upon general verbal ability rather than upon vocabulary or special infor- mation. The reading test was to consist of para- graphs of increasing difficulty, each followed by multiple-choice questions on the content of the paragraph. It was suggested that the ques- tions should measure ability to draw infer- ences, to note detail, and to follow directions. The arithmetical reasoning test was to con- sist of verbally stated problems which involve ability in quantitative thinking. The statement of the problems should be sufficiently simple that reading comprehension is not an impor- tant factor and that the actual reading of the problems consumes a small proportion of the testing time. The problems should be capable of solution by simple arithmetic, and the com- putations should be simple enough that a rela- tively small proportion of the testing time is spent in actually multiplying, dividing, etc. In other words, half or more of the testing time should be spent in thinking out the method of solution. The general classification test which was prepared included 30 sentence completion, 30 opposites, and 40 analogies items, making a total of 100 items. All were multiple-choice items with 5 choices. The reading test was made up of 6 paragraphs with an average of 5 multiple-choice questions on each, making a total of 30 items. The arithmetical reasoning test also contained 30 five-choice items. These numbers were consistent with time limits sug- gested by the Bureau. The items making up each part were arranged in estimated order of difficulty. The attempt was made to make the range of difficulty of the items so great that the easiest item could be answered correctly by approximately 95 per cent of the recruits and the most difficult items could be answered cor- rectly by about 5 per cent. The General Classification Test, Reading RESTRICTED 16 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY Test, and Test of Arithmetical Reasoning are reproduced in the appendix as they were printed by the Navy from copy prepared by Applied Psychology Panel Project N-106. An unclassified account of the preparation of the Basic Battery tests (Form 1) has been pub- lished by the Navy.20 2,4-1 Statistical Analysis of the Basic Battery (Form I) A national sample of 400 answer sheets for each test in the Basic Classification Test Bat- tery (Form 1) was analyzed in detail in order to answer a number of questions concerning the characteristics of the tests making up the battery. The questions are of general interest and can appropriately be asked of any set of tests. Since the answers were specific to Form 1 of the Basic Battery they are not given here in detail. Full information is contained in the original report.14 The specific questions are given below. Internal Characteristics of a Good Test Battery 1. Does each test and subtest yield a suitable distribution of scores? a. Is the range or spread of scores ade- quate ? b. Are the distributions of scores free from significant skewness? c. Are the distributions reasonably simi- lar in shape, so that standard scores for different tests may be considered comparable in meaning? d. Is each test and subtest of suitable dif- ficulty (neither too easy nor too hard)? 2. Are the tests adequately reliable for use in individual placement? 3. Are the correlations among tests suffi- ciently low to provide measurement free from excessive overlapping? 4. Are the correlations among the subtests of the Mechanical Aptitude Test sufficiently high to justify combining the subtest scores into a single Mechanical Aptitude Test score? Similarly, are the correlations among the sub- tests of the General Classification Test suffi- ciently high? 5. Do any of the component subtests of the General Classification Test or the Mechanical Aptitude Test have an unduly large or small influence in determining total-test scores? 6. Do differences exist among the various tests or subtests in the premium granted for speed of performance? Should the time limits for any of the tests or subtests be changed? 7. Concerning the individual items: a. Is there a proper distribution of item difficulty in each test or subtest? More specifically, is the proportion of very easy and very difficult items sufficiently low, and is the dispersion or spread of difficulty satisfactorily adjusted to the average intercorrelation among items in the test or subtest? b. Are the items in each test or subtest arranged in order of difficulty? c. Are the items within each particular test or subtest adequately homoge- neous—i.e., do they succeed reasonably well in measuring the same function? d. Which particular items in each test or subtest should be rejected or revised, because of lack of homogeneity with their fellow items? e. Do the difficult items tend to reduce the homogeneity or internal consistency of any of the tests or subtests of the Basic Battery? Characteristics of Form 1 Tests14 Distribution of Scores. The distributions of test scores for the Form 1 tests were, in gen- eral, satisfactory. Most of the tests were a little too difficult. One, the Surface Development sub- test of the Mechanical Aptitude Test, did not differentiate well among either the very high or the very low men. The easiest questions were not easy enough and the most difficult ones were not hard enough. In consequence, there was a tendency for scores to pile up at both ends of the distribution. Only three of the subtests showed any statistically significant degree of skewness. Likewise three departed from normal in terms of kurtosis; scores on the electrical-pictorial items showed an excessive peaking while the distributions for Surface Development and Block Counting were exces- sively flat. RESTRICTED CONSTRUCTION OF TESTS FOR THE NEW BATTERY 17 Reliabilities. Odd-even reliabilities, corrected by the Spearman-Brown formula, were as shown in Table 5. All reliability coefficients bining these into a single MAT score while re- porting separately the GCT and Reading scores which correlated .81 with each other. Speed. Several of the tests or subtests, par- ticularly Surface Development, but also Sen- tence Completion, Arithmetical Reasoning, and Mechanical Comprehension, put more premium on speed than had originally been in- tended. Revised time limits were recommended. A method of measuring the premium placed on speed in each test is given in Section 15.8. Item Arrangement in Terms of Difficulty. In general the easy items on any test were found in the early part of the test and the hard items toward the end, but in none of the tests was there a smooth and gradual progression from easy to hard items. Rearrangement of the items would have improved every one of the tests in this respect. Item Difficulty and Validity. Information on the difficulty of each item and the biserial cor- relation of each item with the subtest of which it was a part are given in reference 14, The detailed information obtained in this analysis of the Form 1 tests confirmed the judgment of the Bureau of Naval Personnel, which had put the tests into use as soon as they were prepared and without waiting for exten- sive field trials. Original preparation was based on an extensive background of experience in test construction, and it was, therefore, ex- pected that the tests would work well. When the statistical results became available, they fully justified that expectation. Table 5. Reliabilities of Basic Battery tests. General Classification Test .94 Sentence Completion 85 Opposites 84 Analogies 84 .85 Arithmetical Reasoning .83 .97 Block Counting 95 Mechanical Comprehension 84 Surface Development 98 Mechanical Knowledge, Electrical .89 Mechanical Knowledge, Mechanical. . . .90 were based on a random selection of 200 cases from the basic sample of 500. The reliability coefficient for the mechanical aptitude scores was, to some extent, spuriously high, since speed of performance played a considerable role in determining these scores, and it is well known that placing a premium on speed tends to raise the reliability coefficient obtained by the odd-even technique. This was particularly true of the Block Counting subtest of the MAT. Inter correlations. The correlations among the 12 tests or subtests are shown in Table 6. Most of these correlations were satisfactory, but they suggested several possible revisions of tests or procedures. The three intercorrela- tions among the subtests of the MAT were .48, .51, and .60. There is an inconsistency in com- Table 6. Intercorrelations among tests* and subtestsf of Basic Battery (national sample, N = 500). GCT Sent. Comp. Opp. Anal. READ. ARITH. REAS. MECH. APT. Block Count. Mech. Comp. Surf. M.K. Devel. ELECT. M.K. MECH. GCT .89 .90 .90 .81 .69 .60 .46 .57 .48 .53 .49 Sent. Comp. .74 .69 .73 .60 .55 .42 .48 .47 .46 .42 Opposites .71 .73 .61 .47 .38 .46 .36 .48 .46 Analogies .74 .64 .59 .44 .59 .47 .50 .44 READING .69 .56 .44 .52 .47 .51 .46 ARITH. REAS. .61 .52 .51 .51 .47 .41 MECH. APT. .87 .74 .86 .53 .55 Block Count. .51 .60 .42 .45 Mech. Comp. .48 .60 .62 Surf. Devel. .36 .35 M.K., ELECT. .78 M.K., MEGHAN. . in CAPITAL letters. fWritten in small letters. RESTRICTED 18 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY 2 42 The Basic Classification Test Battery, Form 2 When the Bureau of Naval Personnel decided to construct a new form of the Basic Classifica- tion Test Battery, Project N-106 was asked to develop new forms of the three tests which it had originally prepared—the General Classifi- cation Test and the tests of Reading and Arith- metical Reasoning.11 The procedure followed in constructing the Form 2 tests involved four major steps. 1. An experimental edition of each of the three tests was first constructed. An effort was made to match the items of Form 1 in content and difficulty, but in order to allow some selec- tion of items each part of the experimental edition contained about 30 per cent more items than were planned for retention in the final edition. 2. These experimental tests, along with Form 1 of the same tests, were administered to recruits at the U. S. Naval Training Station at Sampson, New York, 3. The items of both forms of each test were statistically evaluated. Information gained in the detailed analysis of Form 1 (see Section 2.4.1) helped here to improve Form 2. 4. Finally, the items to be included in the new Form 2 of each test were selected on the basis of information gained in these item evalu- ations. The individual items in Form 2 of each test were selected on the basis of three criteria: (1) the difficulty of the items in each test; (2) the correlation with total test score; and (3) their comparability with the individual items of the Form 1 tests. By this procedure a second form of each test was constructed which contained items satis- factorily distributed in terms of difficulty, highly correlated with total-test or subtest scores, and comparable to the items in the par- allel Form 1 test. A more detailed description of the proce- dures used in constructing Form 2 of the U. S. Navy General Classification Test and the U. S. Navy Tests of Reading and Arithmetical Reasoning is contained in Chapter 14, which describes the test construction procedures recommended and used by this project. Form 2 of the General Classification Test, Reading Test, and Arithmetical Reasoning Test were put into regular use by the Navy as soon as the Form 2 tests were all completed. The three tests constructed by Project N-106 are reproduced in the appendix. 2 4-3 The Fleet Edition of the General Classification Test Forms 1 and 2 of each of the tests in the Basic Classification Test Battery were de- signed to be administered by competent classifi- cation specialists or experienced officers. Later it appeared desirable to have available a self- administering edition of the General Classifi- cation Test which could be used in the fleet where classification specialists and test scoring machines were not available. Consequently, the General Classification Test, Fleet Edition, was constructed.12 The item content of the Fleet Edition is identical with Form 1 of the GCT. (Form 1 is reproduced in the appendix.) The two forms differ in that the Fleet Edition has modified instructions, an overall time limit instead of separate time limits for the three parts, and does not require a separate answer sheet. Norms for the Fleet Edition and tables for converting raw scores into Navy standard scores (which always have a mean of 50 and a standard deviation of 10) were obtained by giving Forms 1 and 2 to one group of subjects and giving the Fleet Edition and Form 2 to another comparable group of subjects. By this procedure Form 2 served as a bridge connect- ing Form 1 and the Fleet Edition and made possible the establishment of norms and con- version tables. Direct comparison between Form 1 and the Fleet Edition was not possible; since the item content of the two was identical, conversion tables could not be constructed by giving both forms to the same group. 2'4'4 Other Studies of Tests in the Basic Classification Test Battery The Bureau of Naval Personnel requested a statistical analysis of Form 1 of the Mechani- cal Knowledge Test, which had been con- RESTRICTED CONSTRUCTION OF TESTS FOR THE NEW BATTERY 19 structed in the Bureau.9 The data available for analysis consisted of 230 answer sheets for the test, administered in April 1943, at the Bain- bridge Naval Training Station for purposes of preliminary standardization. The Mechanical Knowledge Test consisted of 135 items. Of these, 60 were designed to yield an electrical score and 75 to yield a mechanical score. Both the electrical and me- chanical scores were based on two types of items, pictorial and verbal. The reliability of the total test and the re- liability of the mechanical knowledge portion of the test were both found to be .91, which is satisfactory for individual use. The reliability of the electrical score was found to be .79, hardly satisfactory for individual prediction. All reliability coefficients are odd-even corre- lations corrected by the Spearman-Brown formula. The correlation between electrical and me- chanical scores, .73, was high enough to justify combining the two, instead of reporting sepa- rate scores on the two parts of the test. Differences between the electrical scores and mechanical scores, obtained by the men in the sample studied, were calculated to have a re- liability (odd-even, Spearman-Brown cor- rected) of .44. With such low reliability, dif- ferences between the two scores would have to be relatively large before they could be used with confidence in recommending a man for assignment to a school or billet which required one kind of knowledge (mechanical or elec- trical) to a considerably greater extent than it required the other. It was therefore recom- mended to the Bureau of Naval Personnel that only differences of 11 points or more in Navy standard score should be given serious consid- eration in making such differential assign- ments. Factor Analysis of the Basic Classification Test Battery0 The Basic Classification Test Battery was designed to measure four types of ability: (a) verbal ability, by the General Classification Test and the Reading Test; (b) quantitative reasoning ability, by the Arithmetical Reason- ing Test; (c) mechanical knowledge, by the Mechanical Knowledge Test; and (d) aptitude in dealing with spatial and mechanical rela- tionships, by the Mechanical Aptitude Test. Method of Analysis. The method of multiple factor analysis is appropriate to determine how distinct these four kinds of ability are as meas- ured by the tests of the basic battery. Since sev- eral of the tests consisted of parts which were scored separately, it was possible to obtain 12 different scores, to find the correlations among these 12 scores, and to analyze the intercorre- lations of the 12 scores by factorial methods. The different scores which were used in the factor analysis were: 1. Reading = total score on the Reading Test. Table 7. Reliabilities and intercorrelations of test scores used in factor analysis of the Basic Battery. 1 2 3 4 5 6 7 8 9 10 11 12 1. Reading .82* 2. Opposites .70 .86* 3. Analogies .73 .72 .88* 4. Sent. comp. .68 .70 .69 .85* 5. Arith. reas. .69 .61 .68 .61 .86* 6. Tool rel. .25 .24 .22 .18 .21 .80* 7. Mech. inf. .60 .57 .57 .52 .60 .61 .88* 8. Elect, score .50 .48 .49 .37 .49 . 74f . 84f .84* 9. Mech. score .51 .46 .44 .47 .49 . 79f . 89 f .73 .90* 10. Block count. .31 .23 .32 .30 .43 .22 .34 .27 .31 .90* 11. Mech. comp. .45 .41 .42 .47 .55 .51 .64 .56 .62 .54 .81* 12. Surf, devel. .57 .43 .56 .49 .51 .35 .56 .46 .53 .44 .52 .93* ♦These reliabilities were used for estimating specificity but were not used in fThese correlations were not used in the factor analysis. factor analysis. RESTRICTED 20 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY 2. Opposites = score on the Opposites sub- test of the GCT. 3. Analogies = score on the Analogies sub- test of the GCT. (Factor A) was clearly a verbal factor; the second (Factor B) a mechanical factor; and the third (Factor C) was not well defined in the analysis. These factors were relatively in- dependent of each other, the highest intercorre- lation (that between B and C) being .31 and the lowest (between A and B) being .19. 2. Verbal ability is well represented in the test battery. The Reading Test, the General Classification Test, and the Arithmetical Rea- soning Test all represent this factor. Later validity studies (see Table 11 and, especially, Table 13) showed little consistent differential validity for the three tests in different Service schools. The possibility of pooling scores or of eliminating one of the tests, probably the Read- ing Test, should be considered. 3. An independent quantitative reasoning factor is not represented in the Arithmetical Reasoning Test. Other tests of a quantitative reasoning type might be tried to see if they would give scores which are more differenti- ated from verbal ability and which also meas- ure an ability necessary for success in certain Service schools (see Section 2.5). Table 8. Test projections on the centroid axes. Tests Centroid axes I II Ill 1. Reading + .804 + .294 + .051 2. Opposites + .750 + .364 + .182 3. Analogies + .794 + .337 + .040 4. Sentence Completion + .752 + .328 -.056 5. Arithmetical Reasoning + .785 + .187 -.142 6. Tool relationships + .479 -.464 + .353 7. Mechanical Information + .795 - .231 + .296 8. Electrical score + .700 - .281 + .408 9. Mechanical score + .729 - .349 + .284 10. Block Counting + .517 - .246 -.403 11. Mechanical Comprehension + .725 - . 368 -.169 12. Surface Development + .704 - .126 - .050 4. Sentence completion = score on the Sentence Completion subtest of the GCT. 5. Arithmetical reasoning = score on the Arithmetical Reasoning Test. 6. Tool relationships = score on the pictor- ial section of the MKT. 7. Mechanical information = score on the verbal section of the MKT. 8. Electrical score = score on the electrical sections of the MKT. 9. Mechanical score = score on the mechani- cal sections of the MKT. 10. Block counting = score on the Block Counting subtest of the MAT, 11. Mechanical comprehension = score on the Mechanical Comprehension subtest of the MAT. 12. Surface development = score on the Surface Development subtest of the MAT. Table 7 displays the correlations among these 12 scores. This table was furnished to the proj- ect by the Bureau of Naval Personnel. Factorial Pattern. A centroid factor analy- sis, by the methods devised by Thurstone, gave the factor pattern shown in Table 8. The factor loadings after rotation are shown in Table 9. The major conclusions of the factor analysis study were: 1. Three factors were found. One factor Table 9. Factor loadings after rotation into a simple structure and square root of specificity of tests. Square root of Factor Factor Factor speci- Test A B C ficity 1. Reading + .704f + .078 + .151 + .28 2. Opposites + .758t + .107 -.005 + .36 3. Analogies + .730f O o + + .144 + .36 4. Sentence Completion + .677f -.032 + .217* + .41 5. Arithmetical Reasoning + .565f + .011 + .347* + .44 6. Tool Relationships -.010 + .692f -.003 + .48 7. Mechanical Information + .342* + .592| + .089 + .33 8. Electrical Score + .274* + .674f -.029 + .32 9. Mechanical Score + .208* +.642f + .112 + .41 10. Block Counting + .011 + .038 + .614t + .64 11. Mechanical Comprehension + .088 + .341* + .517f + .35 12. Surface Development + .294* + .257* + .332* + .65 *Factor loadings between .20 and .35. tFactor loadings greater than .50. 4. An independent mechanical ability ap- pears to be best represented by the tool rela- tionships part of the Mechanical Knowledge Test. The two scores now reported, electrical RESTRICTED THE ARITHMETICAL COMPUTATION TEST 21 and mechanical, are very highly correlated and have essentially the same factor composition. If it is desirable to distinguish electrical from mechanical competence, then further work is needed on this differentiation. It may be pos- sible, however, to emphasize the tool relation- ships section of the Mechanical Knowledge Test and so get a score that is relatively inde- pendent of verbal ability. 5. The block counting and mechanical com- prehension sections of the Mechanical Aptitude Test have relatively high loadings on Factor C. The indications from the present battery are not very clear as to the nature of this factor. It may represent speed, spatial reasoning, quantitative reasoning, or some combination of these. A further investigation with additional tests would be necessary to determine what this factor represents. 25 THE ARITHMETICAL COMPU- TATION TEST Construction Correlations between the several tests of the Basic Battery and the factor analysis study just reviewed showed the Arithmetical Reasoning Test to be more closely related to the General Classification Test than was considered desira- ble.6’ 14 The correlations between the individual part scores of the GCT and the Arithmetical Reasoning Test (AR) were all found to be above .60. In view of these findings, a purer test of quantitative ability was requested by the Bu- reau of Naval Personnel. It was desired to have a test which would depend less upon verbal ability than the Arithmetical Reasoning Test apparently did, but which would measure those types of computational skill demanded by Serv- ice school curricula. The Arithmetical Compu- tation Test (AC) was developed to meet this request.13’15 The first step in the development of the new Arithmetical Computation Test was the con- struction of a 60-item, free answer test.13 Prob- lems varied in difficulty from adding 56 and 25 to multiplying 2.96 by .42%. Problems of ad- dition, subtraction, multiplication, and division of whole numbers, fractions, decimals, and mixed numbers were included. The first experimental form of the new test, Form X-l, was administered to 1,430 recruits. The difficulty of each item was determined and used as a basis for selecting the items to be included in the second experimental form, Form X-2. The wrong answers most frequently given in answering the questions on Form X-l were used as distractors in Form X-2. Form X-2 consisted of 50 multiple-choice items. Six, seven, or eight alternative answers were supplied for each item. To prevent the men from choosing an approximate answer without working the problem through to a solution, the alternative “Answer not given” was included for each item; for seven items it was the correct alternative. Form X-2 was administered to over 1,000 recruits at two Naval Training Stations. A sample of 500 papers, half from each station, was analyzed to determine the difficulty and the discriminative value of each item. Form X-2 was found to have a very good distribution of item difficulties. No item was answered correctly by less than 20 per cent or more than 90 per cent of the group. The median was 61 per cent and there was a piling up of items in the 60 to 70 per cent range. This is a desirable feature since it represents a concentration of items where they can make the greatest discrimination. The discriminative value of the items in Form X-2 was very high. The median biserial correlation with total-test score was .695. The Arithmetical Computation Test, Form 1, the final form of the test, was constructed on the basis of the information gained in the analysis of Form X-2, It contains 30 items dis- tributed nearly evenly (6 or 7 each) over the 4 fundamental processes of arithmetic. Nine involve whole numbers; 11, fractions; 6, deci- mals, and 1 both fractions and decimals. There are also 3 problems on percentages. The difficulty range is from 28 to 87 per cent correct for 29 of the items. One very easy item was added at the start of the test. The biserial correlations with total score range from .40 to .84 with a median of .74. Five alternatives were chosen for each item. RESTRICTED 22 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY These were, generally, those most frequently selected by the group taking Form X-2. The distractors used on the final form were those chosen by men who made a lower total score in Form X-2 than the men choosing the correct alternative. “Answer not given” was included as an alternative for each item. The odd-even, and Spearman-Brown cor- rected, reliability of the final 30-item test, esti- mated from the data collected on Form X-2, was .92. Validation A preliminary study13 of the validity of the Arithmetical Computation Test for predicting grades in a number of different types of Service schools was made by correlating the scores of men who took the test at the time of their graduation from Service school with their Serv- ice school grades. The twelve validity coeffi- cients for the Arithmetical Computation Test were found to range from .33 in aviation metal- smith school to .69 in basic engineering school. In six of the ten schools where comparisons could be made, the test had a validity which was as high as, or higher than, any other single Basic Battery test. Such validity coefficients indicate that the Arithmetical Computation Test is a good test if the assumption is correct that the training received in school prior to the test administration does not have a differential effect on performance. Evidence obtained later indicated that this assumption was correct. The evidence also indicated that, although the superiority of the AC over the other tests was not as great as was first reported, the AC showed a higher median validity coefficient (.43) than any test in the battery.15 The Arithmetical Computation Test was found to have high reliability and satisfactory validity. For a number of Service schools it proved to be more valid than any test in the Basic Battery. Yet it failed to achieve the orig- inal purpose of developing a valid test which would show appreciably lower correlations with the verbal tests of the Basic Battery than did the Arithmetical Reasoning Test. The correla- tions of Form X-2 with the GCT and the Arith- metical Reasoning Test for a recruit sample are shown in Table 10, The correlations would un- doubtedly be considerably lower for a sample of Service school men than for this sample of recruits. Nevertheless the correlation with GCT was disappointingly high. Table 10. Intercorrelations of GCT, AR, and AC tests. GCT AR Test (Form 2) (Form 2) AC (Form X-2) .758 .782 GCT (Form 2) .820 In constructing the AC test a very careful attempt was made to prepare a test which would not show a high correlation with verbal tests. The failure of this attempt suggests that it may be impossible to construct a computation test which will not be correlated with verbal measures when a wide range of talent is repre- sented in the sample. 2-5'3 Use of AC Test The new AC Test could be used in any one of four ways: (1) it might be substituted for the AR and thus become part of the Basic Battery; (2) the AR could be retained and scores on both arithmetic tests reported; (3) the AR and AC might both be used, but only the sum of the two scores reported; or (4) the AC could be used as a preliminary screening test and a more difficult arithmetical reasoning test be used as part of a battery of secondary tests. The fourth possibility was recommended to the Bureau of Naval Personnel. If a secondary testing program should be instituted, the AC might well be used as a part of the primary testing battery; and the AR, in a more difficult form, in the secondary battery. Because of its high correlation with GCT, the AC Test could be incorporated into the General Classification Test and thus become a part of a general ability measure. (Chapter 12 contains a discussion of the possible advantages of using both a pre- liminary screening battery of tests and a sec- ond battery for use in making assignments to individual schools or billets.) RESTRICTED GENERAL APPRAISAL OF THE BASIC CLASSIFICATION TEST BATTERY 23 26 GENERAL APPRAISAL OF THE BASIC CLASSIFICATION TEST BATTERY Abilities Measured by the Basic Battery In developing the tests of the Basic Classi- fication Test Battery, the Standards and Cur- riculum Division of the Bureau of Naval Per- sonnel and the Applied Psychology Panel agreed on three major areas of human ability to be in- cluded—verbal ability, mechanical ability, and quantitative ability. The influence of the ex- tensive literature on factor analysis and the identification of the factors which together con- stitute intelligence and learning ability is obvi- ous in this decision. The attempt to measure three reasonably independent factors with the five tests of the Basic Battery was not very successful. The verbal factor was well covered by the GCT and Reading Tests and also by the Arithmetical Reasoning Test. A separate quantitative factor simply did not appear in the factor analysis of the total battery. A mechanical ability factor appeared to be measured by the battery but Usefulness of the Basic Battery Tests In terms of practical usefulness, the new battery worked well. The tests were of satis- factory reliability. Even against grades assigned in the usual not very systematic manner by a diverse group of instructors in class A Service schools, validity coefficients of .50 to .60 were frequently found for individual tests of the battery. Table ll18 shows the mean correlation be- tween each test in Form 1 of the Basic Battery and the final grades achieved by men in 13 types of class A schools. The number of men in each group is also shown. The correlations in the table are those that might be expected if a ran- dom sample of recruits were assigned to the schools. These correlations were obtained by correcting the original r’s for the homogeneity of the sample.17 Mean correlations were com- puted by transmuting each corrected r to z and weighting on the basis of N. Part of the material contained in Table 11, together with a somewhat fuller discussion of the validity data, may be found in reference 21. Table 11. Mean correlations between scores on the Basic Battery tests, Form 1, mentary Service schools (correlations corrected for curtailment in range of talent). and final grades in thirteen ele- Class A school Total number of cases (1) GCT Tests of the Basic Test Battery (2) (3) (4) (5) (6) READ ARI MAT MK(M) MK(E) Best test combination (7) Multi- ple corr. (8) Addi- tive corr. Aviation ordnancemen 184 .63 .59 .59 .54 .35 .42 GCT & ARI .66 .66 Basic engineering 1,480 .52 .52 .63 .52 .46 .39 ARI & MK(M) .66 .63 Diesel (MoMM) 2,160 .42 .35 .36 .26 .43 .46 GCT & MK(E) .53 .51 Electrical 1,747 .52 .52 .59 .44 .35 .49 ARI & MK(E) .63 .63 Fire control man 198 .52 .56 .49 .36 .18 .07 GCT & READ .58 .58 Gunners mate 1,677 .38 .39 .31 .28 .40 .43 READ & MK(E) .49 .47 Hospital corps 449 .50 .45 .36 .27 .22 .30 GCT & READ .51 .51 Machinist mates 755 .33 .27 .44 .48 .48 .45 MAT & MK(E) .54 .53 Radar operators 1,053 .60 .67 .61 .50 .35 .38 READ & ARI .70 .70 Signal 984 .49 .43 .44 .25 .13 .13 GCT & ARI .52 .50 Storekeeper (WR) 678 .44 .47 .59 .37 READ & ARI .60 .58 Torpedoman 880 .32 .35 .28 .27 .39 .35 READ & ARI .44 .43 Yeoman (WR) 738 .62 .59 .63 .38 GCT & ARI .67 .67 less efficiently than was originally hoped. An unplanned factor, not completely clear but pos- sibly measuring ability to handle problems of spatial relations, also appeared in the factor analysis of the battery (see Section 2.4.4). Table 11 also shows the pair of tests which, in combination, served most efficiently to pre- dict grades in each class A school, the multiple correlation based on those two tests, and their additive correlation. The additive correlations RESTRICTED 24 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY were obtained by adding the Navy standard scores on the two tests and correlating their sum with school grades. These additive corre- lations are much easier to compute than are multiple correlations and, for these data, were never more than .03 smaller than the corre- sponding multiple correlation. When the pre- diction tests are highly correlated with each other, it may always be true that additive cor- relations of this type will be as satisfactory as multiple correlations. The method is worth more extensive trial. The tests of the Basic Battery (Form 1) were also found useful in predicting success in more advanced, class B, schools. The correla- tions between the tests and the final average course grades achieved by two samples of men enrolled in the class B Gunners Mate School, Washington Navy Yard, are shown in Table 12.19 effort went into the improvement of Service school grades. Particular attention was paid to developing school examinations which placed greater emphasis on ability to perform and less on verbal memory. One important result of this work shows up in comparing the correlations presented in Tables 11 and 13. In general, the correlations of GOT, Reading, and AR with grades in Service schools giving mechanical and electrical training were lower for Form 2 than for Form 1. In contrast, the correlations of MAT and MKT with grades were higher for Form 2 than for Form 1. This result indicates that the mechanical tests (MAT and MKT) served as better predictors of school grades after those grades were made more dependent on ability to perform and less on verbal knowl- edge. The data summarized in Tables 11, 12, and 13 indicate very satisfactory validity for the Basic Battery tests. These results, the Navy reports, “fully justify the important place the tests have had in the wartime classification program.”18 27 RECOMMENDATIONS FOR FUTURE WORK Future work on the development of tests for the initial classification of military personnel can be recommended along two lines. 2-7-1 Better Criteria Of greatest importance is the development of better criteria for evaluating selection tests. These criteria should be of two types—those from Service schools and those from actual operating performance. The introduction of standardized achievement tests into Service schools (see Applied Psychology Panel, Volume 2, Chapter 17) provides much better school criteria. With improved criteria it becomes pos- sible to determine, much more certainly than one can with poor criteria, which tests are working most efficiently, which should be re- placed or revised, how they should be revised, etc. The general problems of securing useful Table 12. Correlations between Basic Battery test scores, Form 2, and average grades in class B Gunners Mate School, Washington Navy Yard. Correlations with grades Sample 1, Sample 2, Test and subtest N = 84 II o\ C/J General classification test .51 .55 Sentence completion .34 .51 Opposites .33 .42 Analogies .51 .54 Reading .47 .63 Arithmetical reasoning .37 .46 Mechanical aptitude .58 .51 Block counting .37 .38 Surface development .51 .52 Mechanical comprehension .52 .46 Mechanical knowledge (mechanical) .53 .44 Tool relationships .40 .23 Information .50 .47 Mechanical knowledge (electrical) .51 .48 Tool relationships .43 .30 Information .40 .49 Average of reading and mechanical aptitude tests .68 Table 13 gives for Form 2 validity informa- tion similar to that presented in Table 11 for Form 1. The correlations in Tables 11 and 13 have been corrected for the curtailed range of talent in the sample studied.17 During the interval between the introduction of Form 1 and Form 2 of these tests much RESTRICTED RECOMMENDATIONS FOR FUTURE WORK 25 criteria are discussed in Chapter 17 in Volume 2 of this series. Special Ability Tests The success of a military classification pro- gram depends upon the accuracy with which many different types of ability can be meas- picks good radio operators and leaves the men who are good in other respects, but poor in code ability, for assignment to other types of duty. The SOR test (see Chapter 6), the tests for radar operators (Chapter 7), stereoscopic vis- ual tests (Chapter 8), night vision tests (Chap- ter 9), and the Telephone Talker Test (Chapter 10) were all attempts, not always successful attempts, to measure specialized aspects of Table 13. Mean correlations between scores on the Basic Battery tests, Form 2, and final grades tary Service schools (correlations corrected for curtailment in range of talent). in twelve elemen- School GCT Read. Arith. MAT MK(M) MK(E) N No. of school classes Aviation machinist mate .440 .315 .305 .485 .645 .560 750 5 Aviation ordnanceman .535 .390 .350 .285 .355 .490 805 5 Basic engineering .580 .425 .505 .525 .620 .505 1,176 5 Electrical .475 .370 .490 .475 .365 .615 1,062 16 Fire controlman .580 .455 .495 .340 .345 .545 432 5 Gunners mate .380 .325 .310 .495 .560 .545 809 15 Quartermaster .535 .485 .460 .320 .205 .270 470 12 Radioman .270 .255 .270 .165 .195 .235 1,012 16 Storekeeper .590 .415 .500 .265 .095 .310 541 5 Torpedoman .290 .255 .245 .330 .540 .390 786 9 Yeoman .630 .445 .460 .285 .485 .220 857 8 Signalman .430 .343 .406 .360 .245 .345 864 11 ured. Ideally, each man should be assigned, within the limits of military requirements, to the work he can do best. Tests which are gen- eral in nature, e.g., GCT and MAT, are only partially successful in this respect. They do a good job of picking out the generally good men, those who learn easily and who will be success- ful in any one of a variety of jobs. But which school, or which billet, is best for each man depends upon his special combination of abili- ties. Measuring these special abilities requires specialized tests. The Speed of Response Test (SOR) (see Chapter 6) for selecting radio code operators is a good example of the type needed. Good special ability tests will frequently make possible a more equitable distribution of manpower among the various military special- ties than can be accomplished with general tests such as the GCT. The SOR, for example, selects the best prospective radio operators. But since scores on the SOR are not highly cor- related with scores on the GCT, it does not select the men with highest general ability. It man’s ability—special abilities needed for suc- cessful performance in particular military jobs. Most of these tests differ from the GCT and MAT type of test in that they require some kind of performance other than the writing down of answers to questions. In the SOR, one learns three characters of the radio code. In the Tele- phone Talker Test one speaks over a sound- powered phone. Tests of this type are called performance tests. Where they duplicate, in miniature, the work situation for which men are being selected, they are called work-sample tests. Special ability tests are not necessarily per- formance or work-sample tests, but some of the best ones probably will be. Whether a particular ability can be better measured by a perform- ance test or by a paper-and-pencil test is rela- tively unimportant. The reliable measurement of a number of special abilities is the important need. Thorough research studies to develop such tests are strongly recommended. With better criteria to use in validating tests, RESTRICTED 26 THE UNITED STATES NAVY BASIC CLASSIFICATION TEST BATTERY and with better and more diverse measures of specialized abilities, better classification bat- teries can be constructed. It will then be pos- sible to assign each man to the particular job for which he is best suited, with fewer errors and less difficulty than were experienced in World War II. It is important to develop better classification batteries. A future war would be even more technically complicated than was the last one; it would require attaining an even closer ap- proximation to the maximum usage of the best abilities of every single person involved, citizen and soldier alike. The development of these im- provements takes time. We cannot wait, as we did in World War II, for war to break out be- fore starting to improve military classification tests. They must be developed and experience must be gained in their use, during years of peace. RESTRICTED Chapter 3 THE SELECTION OF OFFICERS By Dael Wolfle a Summary would be completed in about 45 minutes. It was required, however, to be a power test, and the candidates were to be allowed approximately an hour and a half to work on the test if they wished. In fulfillment of this request, Project N-106, College Entrance Examination Board, prepared three forms of the U. S. Navy Officer Qualification Test (OQT). Construction The content of the Officer Qualification Test (OQT) was of three kinds. Each was included in an effort to measure an ability important in passing an officers’ training course. Because one essential qualification for success in officers’ training schools is facility in dealing with ver- bal material, a vocabulary test was included. Because a modern Navy is highly mechanized, a mechanical comprehension test was included in which the problems were presented pictori- ally. Because quantitative work and training in mathematics are emphasized in the Navy, the third part of the test consisted of arithmetical reasoning problems that involved no formal knowledge of college mathematics. Part One of the experimental form, on vocabulary, con- tained 60 items; Part Two, on mechanical com- prehension, contained 30 items; Part Three, on arithmetical reasoning, contained 25 problems. The experimental form of the test (called the U. S. Navy Selective Examination, 0-1) was tried out at the New York Office of Naval Offi- cer Procurement (ONOP).1 A sample of 500 papers was obtained for use in item analysis and statistical study designed to aid in the pro- duction of the final form of the test. The odd-even reliabilities of total scores on Form O-l and of scores on its separate parts are shown in Table 1. The intercorrelations of the subtests and the total test are shown in Table 2. The total test and the Opposites subtest both The principal undertaking of the Applied Psychology Panel in the field of officer selec- tion was the development of the U. S. Navy Officer Qualification Test. This test was devel- oped for use by the Offices of Naval Officer Procurement (ONOP’s) in appraising the qual- ifications of men applying for naval commis- sions. After Form 1 was developed, put into use, and subjected to a detailed statistical analysis, Forms 2 and 3 were developed. They were used in the ONOP’s until the end of the war. Several studies dealt with the selection of officers for special types of duty. Tests devel- oped for radar operators were used in selecting tactical radar officers. The NROTC Selection Test was analyzed, and recommendations for its improvement were made to BuPers. The value of various factors in predicting the suc- cess of amphibious flotilla and group command- ers was investigated. A study of combat leadership of infantry officers revealed positive relations of some pre- dictive value with age and with leadership ratings given while at OCS. 31 PREPARATION OF U. S. NAVY OFFICER QUALIFICATION TEST When the Committee on Service Personnel— Selection and Training (which later became the Applied Psychology Panel) was organized in the summer of 1942, the Bureau of Naval Personnel requested a test which could be used to select those officer applicants who had the ability to complete Navy officer training courses satisfactorily. The request specified that the test should be short enough so that it normally a This chapter is based on the work of several of the Applied Psychology Panel projects, chiefly Project N-106. RESTRICTED 27 28 THE SELECTION OF OFFICERS showed satisfactory reliability. Reliability of the Mechanical Comprehension subtest was too low to be satisfactory. For Arithmetical Rea- soning the reliability was borderline. In select- ing items for OQT, Form 1, efforts were made These were reused, and 22 new items were constructed for this subtest. The U. S. Navy Officer Qualification Test, Form 1 is reproduced in the appendix. It con- sists of 100 items divided as follows: 50 oppo- sites, 30 mechanical comprehension, and 20 arithmetical reasoning. It was designed to be self-administered and to be completed within one hour. The test in this form was turned over to the Navy and used at ONOP’s3 until it was replaced by later forms of the test. 31-2 Preparation of Norms A sample of 4,857 tests administered to men at the various ONOP’s during the two-week period ending February 20, 1943, and a similar sample of 5,365 papers from women candidates provided a basis for the preparation of norms for the OQT, Form 1. Separate norms for men and women were furnished to the Bureau of Naval Personnel.4- G 31-3 Statistical Analysis The experimental population used in devel- oping Form 1 of the OQT came entirely from the New York ONOP. Since this regional sam- ple might not fairly represent the whole nation- wide population of officer candidates, an ex- amination of this point was called for. Analysis of Form 1 on a new sample of 500 cases from the New York ONOP and a sample of 500 from various ONOP’s throughout the United States gave evidence that the difficulty values of the separate items remained satisfactorily constant and that they provided a satisfactory range of item difficulty for the national population of officer candidates.7 The biserial correlations between the indi- vidual items and the total scores on the ap- propriate parts of the examination were found to be satisfactorily high for the opposites and arithmetical reasoning portions of the tests. For the mechanical comprehension portion these correlations were somewhat lower on the new sample than they had been on the original sample. In general, however, the correlations Table 1. Reliability coefficients for the Examination, 0-1. Selective Opposites .94 Mechanical Comprehension .73 Arithmetical Reasoning .87 Total test .93 to increase the reliability of both of these latter subtests. The intercorrelations presented in Table 2 were all satisfactorily low except the one be- tween Mechanical Comprehension and Arith- metical Reasoning, and this correlation turned Table 2. Intercorrelations of parts of the tive Examination, 0-1. Selec- Mechanical Compre- Arithmeti- cal Rea- Total hension soning test Opposites .28 .42 .90 Mechanical Comprehension .56 .62 Arithmetical Reasoning .72 out to be substantially reduced in Form 1 of the OQT.7 Items to be included in Form 1 of the OQT were selected on the basis of detailed item analysis data of the type described in Chapter 14. For each item, both difficulty and correla- tion with total score on the appropriate subtest were considered. Fifty items were selected for the Opposites subtest. All had item-test biserial correlations of .45 or higher. The Arithmetical Reasoning subtest was reduced from 25 to 20 items. All those included had item-test correlations of .50 or higher. Form O-l included 18 mechanical compre- hension subtest items which had been borrowed from the Psychological Corporation. It was de- cided not to use these in Form 1. Of the 12 original mechanical comprehension items, only 8 had item-test correlations of .35 or higher. RESTRICTED PREPARATION OF U. S. NAVY OFFICER QUALIFICATION TEST 29 showed only small differences when computed for the United States and the New York sam- ples. The odd-even reliability of the total test, and of its parts, and the correlations between parts remained essentially the same as those deter- mined in the original development of the test. These data are found in Tables 3 and 4. Com- parison of the correlations in Table 3 with those spending figures for the experimental version, 0-1 (compare Table 4 with Table 2). In constructing the OQT, which was to be used throughout the United States, the method used was to administer the test in experimental form to a regional sample, analyze it, and choose items for the standard test on the basis of the analysis. The findings summarized above indicated that this method was satisfactory in the development of Form 1 of the OQT. The method is not, however, recommended for gen- eral use. Ordinarily it will be safer to use a sample of the total population for which the test is intended. 31-4 Preparation of OQT, Forms 2 and 37 The OQT, Form 1 was placed in general use early in 1943, without time for careful prelim- inary standardization and validation. The Bu- reau of Naval Personnel then asked the Applied Psychology Panel to construct two new forms which could be more carefully standardized and which could replace Form 1. The Bureau desired these two forms to be composed of items with similar difficulty distributions and similar distributions of correlations with total- test score. Items from Form 1 might be reused Table 3. Reliabilities of subtests of OQT, Form 1. 2 r r 1 +r N.Y. U.S. N.Y. U.S. Opposites .87 .85 .93 .92 Mech. Comp. .47 .45 .64 .62 Arith. Reas. .76 .75 .86 .86 Total test .87 .82 .93 .90 in Table 1 shows practically no change in the reliability of opposites and total-test scores. The reliability of arithmetical reasoning scores also remained unchanged. An effort had been made to increase this correlation, but at the same time the number of items was reduced from 25 to 20. The result was a test 20 per cent Table 4. Intercorrelations of subtests of OQT Form 1. Opposites Mech. Comp. Arith. Reas. Total test N.Y. u.s. N.Y. U.S. N.Y. U.S. N.Y. U.S. Opposites .23 .16 .42 .36 .91 .90 Mech. Comp. .47 .38 .55 .50 Arith. Reas. .70 . 66 Mean 27.77 25.73 15.98 16.43 10.23 10.42 53.99 52.49 (T 10.83 10.46 3.73 3.73 4.19 4.03 14.82 13.93 shorter and equally reliable. Reliability of me- chanical comprehension scores dropped from an unsatisfactory .73 to an even more unsatis- factory .62, perhaps because of the necessity of including many new and untried items. Correlations between parts of the tests were satisfactorily low, lower in fact than the corre- in Forms 2 and 3. The two new forms were, if possible, to be equated both for men and for women. Forms 2 and 3 were to contain the same number and types of items as Form 1 but were not necessarily to be equivalent to Form 1 in score distribution. An experimental form of the OQT (Form RESTRICTED 30 THE SELECTION OF OFFICERS 0-2), from which items for both Forms 2 and 3 were to be drawn, was developed. This experi- mental form was administered to about 400 officers at the Naval Training School (Indoc- trination) at Dartmouth College, Hanover, New Hampshire, and to about 400 women at the Naval Reserve Midshipmen’s School (WR), Smith College, Northampton, Massachusetts. Detailed item analyses were conducted sepa- rately for the papers from these two groups. Items were selected for OQT, Forms 2 and 3 on the basis of the following criteria:7 1. High correlation between item and total score on the part of the test containing that item, 2. Item difficulties distributed continuously throughout the range of about .05 to .95 with the greatest concentration near the middle of the distribution. 3. Availability of pairs of items having simi- lar difficulty and item-test correlation. One of each pair was included in Form 2, the other in Form 3. The application of these criteria of item se- lection made it possible to construct two forms of the tests which were equivalent in difficulty and (odd-even) reliability both for men and for women. The two tests (OQT, Form 2 and OQT, Form 3) are reproduced in the appendix. 315 Validity7 The validity of the OQT, Forms 2 and 3, was studied by determining the correlations of test scores with training school grades at the Naval Training School (Indoctrination), Dartmouth College, Hanover, New Hampshire, and at the Naval Reserve Midshipmen’s School (WR), Smith College, Northampton, Massachusetts. At Dartmouth grades were available for three courses, seamanship, ordnance, and navi- gation. In addition there was available for each man a rating known as the Officer’s Aptitude Rating. This rating was based on judgments made by the instructors as to each man’s poten- tial officer qualifications. It was possible to compute reliability coefficients for grades in seamanship and ordnance by correlating the sums of grades received in odd-numbered weeks with the sums of grades received in even-num- bered weeks, and then correcting by the Spear- man-Brown formula. The reliabilities varied from .75 to .92, The detailed grades necessary for the computation of a reliability coefficient were not available for navigation nor for the Officer’s Aptitude Rating. Correlations between the three forms of the OQT and the grades for two classes at Dart- mouth are presented in Table 5. The correlation Table 5. Correlations of OQT with grades in NTSch (Indoctrination), Dartmouth College. Class 6 Class i 7 N-. = 360 TV = 403 OQT Final Form Class Sea. Ord. Nav. Apt. average Opp. 1 6 .32 .33 .26 .17 .37 2 7 .37 .25 .35 .02* .36 3 7 .38 .27 .34 .03* .36 Mech. Comp. 1 6 .20 .29 .29 .17 .33 2 7 .28 .28 .36 .10* .35 3 • 7 .32 .30 .38 .12* .38 Arith Reas. 1 6 .39 .29 .52 .10* .49 2 7 .34 .24 .51 .09* .41 3 7 .36 .31 .53 .13 .46 Total test 1 6 .41 .42 .42 .21 .51 2 7 .45 .34 .51 .07* .48 3 7 .47 .37 .50 .09* .50 Mean 6 3.34 3.35 3.19 3.20 3.26 7 3.36 3.38 3.12 3.23 3.27 Or 6 .27 .21 .32 .16 .18 7 .30 .23 .31 .18 .20 *Not significantly different from zero at the 1% level. of about .50 between total-test score and final average grade was high enough to justify using the OQT as an important element in the selec- tion of officer candidates. The consistency of the correlations with final average grades for the two forms of the test and for different classes (r’s of .51, .48, and .50) indicated that the two forms were equivalent and that their usefulness in selecting prospective officers was stable and consistent. At Smith College each student took either Basic Indoctrination and Communications or Basic Indoctrination and Advanced Indoctrina- tion. These three courses included four subjects each as shown in the following outline: RESTRICTED PREPARATION OF U. S. NAVY OFFICER QUALIFICATION TEST 31 Basic Indoctrination: (1) Organization, Naval Law, and Naval Correspondence; (2) Personnel; (3) History and Strategy; (4) Ships and Aircraft. Advanced Indoctrination: (1) Organization and Communications; (2) Personnel; (3) His- made by correlating grades on two of the sub- jects in each course with the other two subjects in that course. For the basic indoctrination course these correlations for three companies were .82, .87, and .87. For communications the correlations for two classes were ,63 and .69. For advanced indoctrination the correlation for one class was .70. The reliabilities of the individual grades were almost certainly as high as these correlations and probably somewhat higher, since one would expect a subject to cor- relate higher with itself than with a somewhat different subject. The correlations of the three forms of the OQT with the grades made by three companies at Smith College are presented in Table 6. The total-test score was found to correlate with the final weighted average grade to the extent of .33 to .40 for those who took the communica- tions course and to the extent of .39 to .47 for those who took advanced indoctrination. The test was therefore shown to work less satisfac- torily for women than for men. In view of the differences in validity of the test as applied to the Dartmouth and Smith College populations it was suggested to the Bureau of Personnel that it would be preferable to use Forms 2 and 3 of the OQT for men and to develop a different test for women officer candidates. 3-1,6 Statistical Analysis After Forms 2 and 3 of the OQT had been put into use, answer sheets for a sample group of candidates were collected from the ONOP’s. Item difficulties, biserial correlations of each item with the score on the appropriate subtest, and reliabilities and intercorrelations of the subtests were determined.10 On the basis of this statistical information, the following recommendations were made to the Bureau of Naval Personnel: 1. Whenever practicable, experimental tests should be tried out on a sample exactly similar to the group for whom the regular form of the test was intended. It would have been bet- ter to have tried out the experimental edition of Forms 2 and 3 on a group of applicants rather than on a group of already successful Table 6. Correlations of OQT with grades at Naval Reserve Midshipmen’s School (WR), Smith College. Company 1 Company 2 Company 6 N —152 N-- = 160 A'' =115 Final OQT Adv. weighted Form Co. Indoc. Comm. Indoc. average 1 .33 .08* .20* 2 2 .40 .24 .31 6 .27 .30 .30 Opposites 1 .34 .10* .21 3 2 .39 .31 .36 6 .33 .34 .36 1 .23 .30 .31 2 2 .12* .13* .14* Mechanical 6 .19* .08* .15* Comprehension 1 .19* .29 .28 3 2 .04* .12* .11* 6 .23 .13* .20* 1 .26 .27 .29 2 2 .23 .29 .29 Arithmetical 6 .43 .27 .38 Reasoning 1 .32 .31 .34 3 2 .29 .27 .30 6 .48 .36 .45 1 .40 .24 .33 2 2 .41 .31 .37 6 .39 .34 .39 Total test 1 .42 .24 .34 3 2 .41 .36 .40 6 .47 .41 .47 1 3.30 3.16 3.21 Mean 2 3.33 3.19 3.23 6 3.25 3.27 3.26 1 .24 .21 .20 a 2 .25 .21 .21 6 .25 .21 .22 *Not significantly different from zero at the 1% level. tory and Strategy, and Correspondence and Filing; (4) Ships and Aircraft. Communications: (1) Naval Communications (General); (2) Fundamentals of Radio; (3) Codes and Ciphers; (4) Touch Typewriting. Estimated reliabilities of the grades were RESTRICTED 32 THE SELECTION OF OFFICERS candidates. The men in the school group dif- fered slightly in average ability from the officer candidates at the ONOP’s and showed a smaller range of ability. 2. An attempt to increase the reliability of the mechanical comprehension subtest of the OQT was recommended. Its reliability of .74 to .78 was too low to be acceptable. 3. The three subtests were sufficiently inde- pendent to justify reporting separate scores rather than one total score. Forms 2 and 3 of the OQT continued in use at the ONOP’s until the end of the war. 3 2 OTHER STUDIES OF OFFICER SELECTION 3-21 The Prediction of Combat Leadership In fulfillment of a request from the Office of the Adjutant General, War Department, the Applied Psychology Panel conducted a pre- liminary study of combat leadership. Two classes of infantry officer candidates and two classes of field artillery officer candidates were studied. For each of the four classes, correla- tions were computed between graduation or failure to graduate and a number of other vari- ables which might be used to predict officer success. In addition, combat ratings of 176 in- fantry company officers were compared with several types of officer candidate school (OCS) records.9 Prediction of Success in Infantry OCS Information was obtained on 40 items con- cerning the men in two classes at Fort Hen- ning. Many of these items were found to be completely unrelated to success or failure in OCS. Those which appeared to be of greatest possible usefulness are given in Table 7 to- gether with their correlations with OCS suc- cess or failure. Some of the correlations presented in this table are at variance with common belief, for example, the finding that hobbies, classified as vigorous versus sedentary, showed a zero correlation with the criterion of being commissioned or not being commis- sioned. Table 7. Correlations of seven items obtained be- fore or upon entrance to infantry OCS against the criterion of being commissioned or not being com- missioned. Criterion vs following Class 304 Class 311 items* N = N = about 200 about 200 1. Average first and sec- ond AGCT scores .45 .46 2. Amount of schooling .39 .07 3. Age at entrance to OCS —.12 —.04 4. Father’s occupation (leadership vs non- leadership) .12 .32 5. Athletic participation —.08 —.10 6. Hobbies (hobbies clas- sifted as vigorous and athletic vs mild and sedentary) .00 .02 7. Leadership as indicated by former civilian po- sitions .20 .40 * Correlations 1, 2, 3 are biserial; 4, 5, 6, 7 are tetrachoric. At this OCS, both platoon leaders and fellow officer candidates made periodic ratings or rankings of the leadership qualities of the members of their platoon. These ratings were Table 8. Biserial correlations of leadership ratings against criterion of being commissioned or not being commissioned in infantry OCS. Criterion vs following items Class 304 Class 311 Platoon leaders’ ratings at the end of 5 weeks Platoon leaders’ ratings at the end of 10 weeks Platoon leaders’ ratings at the end of 15 weeks Final company ratings .36 (N= 201) .70 (2V = 143) .85 (N = 145) .87 (N= 145) .84 {N — 204) * .78 (N — 143) .84 (N = 126) .90 {N= 126) * This high r was apparently due to the large number of ROTC candidates in Class 311. These men had received considerable mili- tary training and their leadership qualities were apparently evaluated quickly by their platoon leaders. considered in deciding which candidates should be commissioned. Some of the correla- tions of the ratings with the criterion are pre- sented in Table 8. Table 9 shows the intercorrelations among selected items of information obtained. None of the items obtainable before or upon entrance to OCS was highly predictive of later RESTRICTED OTHER STUDIES OF OFFICER SELECTION 33 performance in school. The highest correlation was with the Army General Classification Test score. The next highest was with leadership as indicated by former civilian positions. This measure looks hopeful for selection purposes. Attempts to measure the leadership actually Table 10. Biserial correlations of nine items obtained before or upon entrance to field artillery OCS against the criterion of being commissioned or not being commissioned. OC 80 OC 82 Criterion vs following items N = about 200 N = about 200 Army General Classifi- cation Test .29 .58 Education (number years in school beyond elementary level) .33 .36 Age —.22 —.67 Participation in athletics —.20 .16 Civilian occupation (rating from Barr Scale) .17 .32 Length of military service —.52 .74 Mathematical test (given first week of course) .48 .40 Mathematical courses (number previously taken in school) .46 .00 Administration and military law (test given on arrival) .39 .12 Table 9. Intercorrelations among selected items at infantry OCS. Class 304 Class 311 First AGCT and sec- ond AGCT .64 (iV = 213) .72 (N = 212) Platoon leaders’ ratings for leadership, end of 5th week, and aver- age AGCT .21 {N= 200) .20 (N— 202) Platoon leaders’ ratings for leadership, end of 5th week, and high- est military rank at- tained in college ROTC —.16 (iV = 37) .15 (N = 119) Platoon leaders’ ratings for leadership, end of 5th week, and end of 15 th week .65 (N= 99) .49 (W= 114) shown in civilian positions would be well worth while. The information gained might materi- ally improve the selection of officer candidates. Prediction of Success in Field Artillery OCS Data from the Field Artillery Officer Candi- date School at Fort Sill are given in Tables 10 and 11. Background information available upon en- trance to OCS was generally more useful in predicting success of field artillery officer can- didates than it was for infantry officer candi- dates. Previous education and grades on mathematics examination scores both showed useful relationships (correlations of .33 to .48) with success in field artillery OCS. The differences in the correlations found for classes 80 and 82, which sometimes were very large, appeared to be due to differences in the men enrolled in the two classes. Class 80 con- sisted of men from all branches of Service and included a number who had been in the Army for a considerable period. Class 82 consisted almost entirely of ROTC candidates. The dif- ferences in correlations for the two classes Table 11. Biserial correlations of ratings of leader- ship against the criterion of being commissioned or not being commissioned in field artillery OCS. OC 80 OC 82 Criterion vs following items N = : about 200 N = about 200 Tactical officers’ ratings at end of 5th week (leadership, courage, personal appearance) .80 .74 Tactical officers’ final ratings before grad- uation (leadership, courage, personal ap- pearance) .90 .80 Gunnery officers’ rat- ings. First rating during 5 weeks’ course (ability shown in firing problems) .52 .86 Gunnery officers’ rat- ings. Second rating during 5 weeks’ course .88 .76 Ratings by classmates at end of 5th week .54 .70 Ratings by classmates just before gradua- tion .96 .90 RESTRICTED 34 THE SELECTION OF OFFICERS made prediction hazardous, but, in general, it appeared probable that success in field artillery OCS could be predicted more accurately from data available before entrance than was the case for infantry OCS. The agreement between combat ratings and the leadership ratings assigned while in school at Fort Benning can be seen in Figure 1. All the men who received school ratings in the first (top) fifth for leadership received combat rat- ings of “Satisfactory” or better. The percent- ages of combat ratings of “Superior” and “Excellent” became steadily smaller for the groups which received lower leadership ratings in school. There was no significant relation between combat success and AGCT score. But the group was a select one to begin with, the average AGCT score being 124.6. There was a fair relation between age and combat rating. The superior and excellent offi- cers on the whole came from the age group 22 to 28. Proportionately fewer men below 22 or above 28 were so rated. After these men graduated, the academic re- quirements were changed. The effect which the changed standards would have had on these 176 officers is shown below. Per cent who Number who would not would not have gradu- have gradu- ated under ated under changed changed requirements requirements Superior group 4.3 1 Excellent group 10.3 9 Very Satisfactory group 24.3 10 Satisfactory group 29.4 5 Unsatisfactory group 50.0 4 Figure 1. Agreement between combat ratings and ratings of leadership qualities shown in infantry OCS. The trend of percentages shown by these fig- ures is very encouraging. The actual figures on which they are based, however, show that the gain of graduating four fewer officers who later received unsatisfactory combat ratings would have been made at the expense of reject- ing 20 men who later received combat ratings of “Very Satisfactory,” “Excellent,” or “Su- perior.” When officer candidates are plentiful this expense is justified. Conclusion Regarding Combat Leadership The results of this brief study of combat leadership were encouraging enough to war- rant further investigations in this field. Cer- tainly the importance of accurate prediction of successful combat leadership is great enough to The Relations between Success in Infantry OCS and Combat Ratings Ratings were obtained on the combat effi- ciency of 176 company officers all of whom had graduated from the infantry OCS at Fort Benning. These ratings were made on a five- point basis by the regimental commanders or their executive officers and by the battalion commanders. The ratings were distributed as follows: Rating Number Per cent Superior 23 13.1 Excellent 87 49.4 Very Satisfactory 41 23.3 Satisfactory 17 9.7 Unsatisfactory 8 4.5 RESTRICTED OTHER STUDIES OF OFFICER SELECTION 35 justify a good deal of effort directed toward improving the methods used to select company officers. Of first importance in such a study will be the task of obtaining careful, detailed, and reliable measures of success or failure on the part of a large number of officers engaged in a variety of military duties. These measures are necessary as a criterion to use in evaluating possible selection tests or measures. 32 2 NROTC Selection Test2 The Bureau of Naval Personnel requested the Applied Psychology Panel to make an analysis of the NROTC Selective Examination and to suggest revision of that test. The test consisted of four parts: 100 multiplication problems; 50 block counting items; 50 vocabu- lary items; and 50 arithmetical reasoning items. The study of the NROTC Selective Examina- tion included an item analysis of a sample of 500 papers, a study of the reliability and inter- correlations of the parts of the test, and a study of the validity of the test. Grades in deck offi- cers school (N = 216) and engineering school {N = 193) were used as criteria. This investi- gation led to the conclusion that the total score could be made somewhat better for predicting grades by omitting the multiplication section and by simply summing the scores on the other three parts instead of using the complicated derived-score system previously in use. For two groups of students, correlations between total derived scores on the test and final grades were .49 and .51. Using unit weights for the three tests (Block Counting, Vocabulary, and Arithmetical Reasoning) gave multiple validity coefficients of .50 and .55 for the same two groups. Omitting the Block Counting test and weighting the other two equally reduced these correlations to .45 and .49. It therefore seemed probable that the block counting section could also be eliminated and the vocabulary and arithmetical reasoning sections improved suf- ficiently to give validity coefficients as good as those found for all three tests. Detailed sug- gestions for revising these two subtests were made to the Bureau of Naval Personnel.2 Selection of Flotilla and Group Commanders11 Project N-117 of the Applied Psychology Panel conducted a brief investigation of the in- formation which might be useful in selecting flotilla and group commanders in the Amphibi- ous Training Command, U. S. Atlantic Fleet. The study was made at the request of that Command. Nine possible predictors were studied in relation to fitness report data (Items 8, 10, and 13 of the Report on the Fitness of Officers, NavPers 310) for a group of 76 officers, 42 of whom were graduates of the Naval Academy and 34 of whom were not. The predictors included in the study were OQT, Form 3; MKE, Form 2 and MKM, Form 2 (see Chapter 2) ; Officer Personnel Inventory (see Chapter 4) ; age, rank, months of sea duty, whether an academy graduate or not, and in- terviewer’s overall evaluation. The interviewer’s evaluation showed higher correlations (.40, .36, and .45) with the three fitness report items than did any of the other predictors. The small number of cases included in the study and the fact that both the inter- viewer’s opinion and the fitness reports were, in part, based upon the same factors and the same kind of general opinion—both require that these results be interpreted with caution. 324 Selection of Tactical Radar Officers Two projects of the Applied Psychology Panel contributed significantly to the develop- ment of tests and procedures used by the Navy to select officers for duty in combat informa- tion centers. These tests are described in Chap- ter 7, “Selecting Radar Operators.” RESTRICTED Chapter 4 ELIMINATING THE EMOTIONALLY UNFIT By Dael Wolfle a Summary unstable and likely to break down under the stress of hazardous duty. Physical examina- tions and aptitude tests are not designed to pick out this type of person, and they do not do it well. Brief psychiatric examinations can some- times be used to identify the emotionally un- stable or psychiatric cases. But the limited number of qualified military psychiatrists and the large number of men to be examined com- bined, during World War II, to make it impos- sible to give thorough psychiatric examinations to all men for whom such examinations would have been desirable. Civilian experience prior to the war led to the belief that it would be valuable to develop a test for preliminary screening of psychiatri- cally undesirable men in the military Services. It was agreed that an inventory should be de- veloped which would emphasize biographical items. A contract to develop such an inventory was written with Brown University. The proj- ect started under Section D-2 of NDRC, con- tinued under the National Research Council Committee on Service Personnel—Selection and Training, and was completed under the Applied Psychology Panel as Project N-113. The result of this work was the Personal Inventory. In many Service situations it proved to be a highly useful device for the preliminary screening of large groups of men. Those show- ing evidence of emotional disturbance or insta- bility could be interviewed by the psychiatrist while the others could be passed without inter- view. Where regulations required interviewing every man, more time could be spent with those whose scores indicated considerable likelihood of emotional instability. 42 THE PERSONAL INVENTORY Four forms of the Personal Inventory were developed. Three were designed for use with enlisted men and one for use with officers. The personal inventory was developed in order to have available a device for making quick identification of emotionally unstable men who are likely to break down under the stress of hazardous duty. The Personal Inventory consists of paired statements. From each pair, the man being tested selects the one which best fits him. Each pair includes a symptomatic and a nonsympto- matic (normal) statement. The number of symptomatic statements checked constitutes a man’s score. With many different Service groups it was demonstrated that men making high scores were much more likely to be judged unfit for service, when examined by a psychiatrist, than were men making low scores. The Personal In- ventory was found to discriminate better than other tests and methods tried out between ac- ceptable men and those given psychiatric dis- charges. The Personal Inventory was used as a pre- liminary screening device to aid the psychia- trist by identifying the men most likely to be classed as emotionally unfit. It was thus pos- sible for the psychiatrist to interview only a fraction of the men, but to find, in that frac- tion, most of those who would have been given psychiatric discharges if all men had been in- terviewed. 41 INTRODUCTION In selecting men for special types of duty, such as officer candidate schools, submarines, parachute troops, etc., physical examinations weed out the obviously physically unfit, and aptitude tests weed out most of those too inapt to learn their special duties. A group is left who are disqualified because they are emotionally a This chapter is based primarily upon the work of NDRC Project N-113. 36 RESTRICTED THE PERSONAL INVENTORY 37 4-2-1 Enlisted Men's Personal Inventory, Long Form1 Development The original form of the Personal Inventory (PI) for enlisted men contained 145 items, most of which were in a forced-choice form. In answering each item one must choose between two statements. One alternative in each item presents a symptomatic attitude or type of be- havior; the other a nonsymptomatic attitude or type of behavior of approximately equal acceptability. For example: other symptoms which occurred significantly more frequently in one group than in the other. Analysis of the case histories was sup- plemented by conferences with naval psychia- trists and other medical officers and by reference to naval and military psychiatric literature. The 300 items originally constructed were sifted to give appropriate emphasis to each area to be covered in the final inventory. Those finally selected were then further refined to make them understandable and relatively in- offensive to the subjects and to make the two alternatives in each item approximately equally desirable. In most items, however, there was a slightly greater apparent desira- bility for the choice usually made by normal individuals, the nonsymptomatic alternative. The inventory was constructed so that the answers could be recorded on a separate an- swer sheet which could be scored either by hand or by machine. The inventory was printed in two formats. In Format A the answers were recorded through holes in the test booklet. The scoring key collected together the answers to related items. This feature made it possible to secure subscores on each of 20 separate clusters of items, for example, hypochondriasis, sociability, or mood swing. Format B did not have this clustering principle or the mechanical elabo- rateness which it required. In Format B only a single score was obtained. The two formats were shown experimentally to be equivalent.8 In practice, Format B soon became widely used and Format A practically forgotten. Since the items on the two formats were identical, further discussion of the long form will not dis- tinguish between Formats A and B. In prac- tically all cases Format B was the one used. That form is reproduced in the appendix. A later study19 from the Medical Research Department, U. S. Submarine Base, New Lon- don, Connecticut, supplied statistical support for the general preference for Format B over Format A, This study presented the inter- correlations of 58 of the 60 scored items. Of a total of 1,653 (tetrachoric) correlations, 1,220 fell within the range —.10 to +.19. A detailed factor analysis of the data was not made, but I have felt bad more from head colds ( ) I have felt bad more from dizziness ( ) For several reasons the items were phrased in this forced-choice form rather than in the more familiar Yes-No or Yes-No-Doubtful form. The most important reason was the be- lief that this form would elicit more honest answers than a Yes-No question on the same symptom, e.g.: I frequently feel dizzy Yes No In addition, statistical treatment of the an- swers for the form used was relatively easy, and it was believed less subject to a “halo” effect than the Yes-No form of item. One dis- advantage of the form was the fact that many men objected to being forced to choose between two statements neither of which applied par- ticularly well to them. In spite of this objection, the method worked very effectively. The items were constructed on the basis of a careful study of the case histories of a group of normal enlisted men and another group of psychiatric cases. The case histories were ob- tained from an OSRD report entitled “Analy- sis of 100 Psychiatric Cases from the Chelsea Naval Hospital.” All case records were ex- amined for biographical items which appeared to differentiate the normal from the psychiatric histories. The term “biographical” was con- strued broadly to include educational, occu- pational, social, and attitudinal factors and any RESTRICTED 38 ELIMINATING THE EMOTIONALLY UNFIT Table 1. Item analysis of the Personal Inventory, long form. N answering Percentage answering N answering Percentage answering AT answering item given alternative given alternative N answering item given alternative given alternative Item Nor- Dis- Nor- Dis- Nor- Dis- Critical Item Nor- Dis- Nor- Dis- Nor- Dis- Critical No. Choice mals charges mals charges mals charges ratio No. Choice mals charges mals charges mals charges ratio 1 R 1,001 384 190 113 19.0 29.4 4.3 79* L 990 374 181 118 18.3 31.6 5.2 2 R 994 380 796 306 80.1 80.5 0.4 80* L 998 383 65 108 6.5 28.2 8,6* 3 R 980 360 814 302 83.1 83 .9 0.4 81 L 1,003 385 456 194 45.5 50.4 1 .4 4* R 956 362 130 121 13.6 33 .4 7.0*OS) 82* L 1,001 385 102 195 10.2 50.6 15.4*05') 5* R 994 381 179 116 18.0 30.4 4.6 83 R 1,002 378 823 326 82.1 86.2 1 .8 6 R 1,000 379 81 56 8.1 14,8 3.5 84 R 999 383 261 202 26.1 52.7 7.9* 7 R 995 380 358 164 36.0 43.2 2.4 85 L 1,001 385 218 92 21 .8 23.9 0,8 8* R 1,003 385 37 41 3.7 10.6 4.1 86 L 1,002 384 461 198 46.0 51 .6 2.0 9 R 1.004 385 93 38 9.3 9.9 0.6 87 R 1,000 384 565 266 56,5 69.3 4.2 10 R 994 384 61 26 6.1 6.8 0.7 88 L 994 382 526 214 52.9 56.0 1 .0 11* R 998 384 139 120 13.9 31 ,3 6.5* 89 L 1,001 385 61 49 6.1 12.7 3.7 12 L 996 383 161 128 16.2 33.4 6.3* 90* L 991 378 45 48 4.5 12.7 4.4 13* R 1,000 384 67 50 6.7 13.0 3.2 91* R 999 382 9 71 0.9 18.6 8.9* 14 L 1,001 385 250 118 25 .0 30.6 2.2 92 R 1,001 382 142 88 14.2 23.0 3.7 15 L 1,001 385 190 120 19.0 31 .2 4.5 93 L 999 383 118 70 11 .8 18.3 2.7 16* R 994 384 309 192 31 1 50.0 5.5* 94* L 1,003 385 127 144 12.7 37.4 9.5* 17 R 1,001 381 187 105 18.7 27.6 3.5 95 L 1,002 383 199 99 19.9 25.8 2.3 18 R 990 380 120 70 12.1 18.4 2.7 96* L 1,001 384 22 94 2.2 24 5 9.8 *(S) 19 L 999 384 107 88 10.7 22.9 5.2 97* L 996 384 486 241 48.8 62 .8 4.7 20 R 1,000 384 119 72 11 .9 18.8 3.1 98* R 1,001 383 163 163 16.3 42.6 9.8* 21 R 1,001 383 133 103 13.3 26.9 5.5 99 L 1,002 381 371 151 37 0 39.6 1 .0 22 R 996 384 503 231 50.5 60.2 3.0 100* L 1,003 384 158 117 15.8 30 5 5.4 23 L 994 382 184 108 18.5 28.3 3,5 101 L 999 385 289 165 28,9 42.9 4.9 24 L 992 382 262 118 26.4 30.9 1 .8 102* R 1,003 384 152 139 15.2 36.2 8.0* 25 L 1,001 385 93 77 9.3 20.0 5.0 103* L 1,003 384 534 298 53.2 77.6 9,5 *(S) 26 L 1,003 385 356 176 35.5 45.7 3.4 104 R 1,004 383 183 126 18.2 32.9 5.6* 27 L 996 384 131 93 13.2 24.2 3.7 105* R 998 384 402 242 40.3 63 0 7.7* 28 L 1,001 384 25 26 2 .5 6.8 9.1* 106 R 1,000 384 260 145 26.0 37.8 4.2 29 L 1,003 385 269 149 26.8 38.7 4.2 107* L 1,002 385 50 133 5.0 34.5 12.0*(5) 30* R 997 383 274 193 27 .5 50.4 6.5* 108* L 997 382 343 242 34.4 63.4 9.9* 31 R 999 381 191 161 19.1 42.3 9.1*(S) 109* R 999 381 16 60 1 .6 15.7 7,2* 32* L 999 379 331 223 33.1 58.8 8.9* 110* R 1,000 384 26 100 2 .6 26.0 10.2*(S) 33 R 998 384 694 290 69.5 75 .5 2.6 111 R 996 382 338 130 33.9 34.0 0.0 34 R 988 376 60 57 6.1 15.2 4.6 112* R 986 378 30 65 3.0 17.2 7.1* 35* L 998 374 530 275 53.1 73.5 7,5* 113* L 1,002 384 45 92 4.5 24.0 8.2* 36* R 997 367 34 143 3 .4 39.0 14.1*(5) 114 R 996 379 98 65 9,8 17.2 3.3 37 R 1,000 378 461 214 46.1 56.6 2.8 115* L 1,002 383 39 65 3.9 17.0 6 5* 38 L 998 385 407 224 40.8 58.2 5.7* 116 R 1,001 385 56 30 5 .6 7.8 1 ,2 39* R 1,003 385 53 98 5.3 25 5 8 7*(S) 117 L 988 382 529 243 53 .5 63.6 3.4 40 L 1,004 385 466 182 46.4 47.3 0 3 118 L 934 370 112 54 12.0 14.6 1 .4 41 L 1,002 385 424 219 42,3 56.9 5.1 119 L 1.000 382 299 189 29.9 49.5 5.6* 42* R 1,003 383 119 158 11 9 41 3 10.8*15) 120 L 1,003 384 61 40 6.1 10.4 2.4 43 L 1,000 381 700 318 70.0 83.5 3 0 121* L 996 382 209 188 21 .0 49.2 8.3* 44 L 999 383 214 103 21 4 26.9 2.3 122* L 1,003 383 19 76 1 ,9 19.8 8.8* 45 L 1,002 383 103 67 10.3 17.5 3.7 123* R 996 382 26 120 2.6 31 4 11 9*(5) 46 R 1,002 383 873 342 87.1 89.3 1 0 124 L 998 380 633 266 63.4 70.0 2.5 47* L 999 385 12 157 1 .2 40.8 15.9*(5) 125 R 991 382 20 12 2.0 3.1 1 .0 48* R 994 380 93 53 9,4 13 .9 2.4 126* L 992 379 89 83 9.0 21 .9 5.7* 49 R 969 370 585 231 60.4 62 .4 0.3 127 L 1,000 383 244 123 24.4 32.1 2.9 50* R 986 380 115 98 11.7 25.8 5 .6* 128 L 1,001 385 157 87 15.7 22.6 2.9 51* L 1,001 384 144 211 14.4 54.9 15.0*(5) 129* L 1.001 385 109 135 10.9 35.1 9.2* 52* L 1,000 382 10 51 1 0 13.4 6.9* 130 L 996 384 234 137 23.5 35.7 4.3 S3 R 1,002 385 275 194 27.4 50.4 7.7* 131 L 1,000 381 122 94 12.2 24.7 5.4 54 L 1,002 383 533 269 53.2 70.2 6.1* 132 L 1,000 384 156 98 15.6 25.5 4.0 55 L 1,000 384 517 257 517 66.9 5.2 133 L 995 378 98 48 9.8 12.7 1 .5 56 L 998 380 337 175 33.8 46.1 3.5 134 R 998 381 26 11 2.6 2.9 0.0 57 L 994 385 106 58 10.7 15 1 1 .8 135 L 998 383 152 72 15.2 18.8 1 8 58 R 1,002 385 317 133 31 .6 34.5 1 .0 136 L 997 385 228 160 22 .9 41 .6 6.7* 59* L 1,001 382 143 109 14 3 28.5 5 .9* 137* L 998 384 95 118 9.5 30.7 8.2*(S) 60 L 999 384 149 71 14.9 18.5 1 8 138* R 997 385 364 243 36.5 63.1 9.3* 61* L 999 380 214 171 21 .4 45.0 7.1*(5) 139* L 1,003 383 22 54 2.2 14.1 6.5* 62* R 992 377 477 228 48.1 60.5 4.0 140* L 1,001 384 334 195 33.4 50.8 5.2 63* L 972 366 196 197 20.2 53.8 9.8*(5) 141* R 998 384 14 52 1 .4 13.5 7.1* 64 K 1,000 379 695 269 69 5 71 0 0.7 142* L 1,001 385 50 75 5 .0 19.5 7.1* 65* R 999 381 119 142 11 9 37.3 5.6* (S) 143* L 1,001 383 10 31 1 .0 8.1 4.9 66* R 998 377 41 109 4 I 28.9 10.5*(S) 144 R 999 384 74 29 7.4 7.6 0.6 67 R 998 384 294 147 29.5 38.3 3.2 145 R 1,000 385 139 79 13.9 20.5 3.0 68 L 993 383 262 135 26.4 35.2 2.9 69* L 995 381 96 152 9.6 39.9 11,5*(5) Note—The item numbers are shown in column 1. The 60 items marked 70* L 1,002 383 217 165 21.7 43.1 7.5* with an asterisk are the 60 most discriminative items found in the first 71 72 R R 997 994 381 382 182 104 99 87 18.3 10.5 26.0 22 8 3.2 5 0 item analysis 1 The 60 items marked with an asterisk in the last column are 73 L 998 382 341 168 34.2 44.0 3.4 the 60 most discriminative items found in the second item analysis.5 The 74* L 997 381 485 263 48.6 69.0 6.9* 20 critical ratios in the final column marked with the letter S) indicate 75* L 1,000 384 180 152 18.0 39.6 7.9*(5) the 20 items which later became the short form of the Personal Inventory.7 77 L 1,004 384 21 I 100 21 .0 26.0 2.0 The choice indicated in the second column is the choice marked more 78* L 994 380 247 199 24.8 52.4 7.9* frequently by the discharge group than by the normal group. RESTRICTED THE PERSONAL INVENTORY 39 inspection seemed to indicate the presence of two factors. There was no inspectional support for the use of the twenty separate cluster scores obtained with Format A. Correlations between answers to 37 PI items and Otis In- telligence Test scores are also contained in reference 19. Scoring Key A scoring key1 was devised by administering the long form to a group of 1,210 newly enlisted men who had been passed upon favorably in a psychiatric interview, and to 105 men who had been passed upon by a discharge board and were being discharged for psychiatric reasons. These men were tested while in the psychiatric ward. Both groups were obtained at the U. S. Naval Training Station, Newport, Rhode Island. Papers were excluded for all men who omitted more than five items and for those who had at some time been under psychiatric obser- vation but had later been returned to their training companies for trial duty. These exclu- sions left a group of 1,004 normals and a group of 84 psychiatric discharges. An item analysis of the answers of these two groups to each of the 145 items was made. Dif- ferentiation of each item between the two groups was measured in terms of the critical ratio. Sixty items were found to have critical ratios of 2.7 or greater. The scoring key for the long form of the PI is based on these 60 items. A man is given one point for each of the 60 items to which his answer is the same as that characteristically given by the discharge group. Low scores are, therefore, “good” scores. A second analysis5 of the same type was made on a new group of 385 psychiatric dis- charges from Newport Naval Training Station and the same group of 1,004 Newport normals. Forty-nine of the 60 items having critical ratios of 2.7 or greater in the first analysis were among the 50 most discriminative items in the second analysis. The detailed results of this second item analysis are given in Table 1. Comparison of the ability of the PI to distin- guish between normal and psychiatric groups showed only insignificant differences between the two scoring keys. The original scoring key was therefore retained. The distribution of critical ratios of the 60 scored items is given in Table 2. The median critical ratio was over 4.0, and all were greater than 2.7. Table 2. Critical ratios Inventory scoring key. of items in the Personal C. R. (diff. /a d.ff ) Number of items 2.7-2.9 4 3.0-3.9 24 4.0-4.9 17 5.0-5.9 10 6.0-6.9 3 7.0-7.9 2 60 Reliability The reliability of scores on the long form was determined on four groups of subjects. The data are given in Table 3. Table 3. Reliability coefficients man-Brown corrected) for the Personal Inventory. i (odd-even, Spear- long form of the Group Mean o' rt 508 Newport recruits 610 amphibious forces men 188 Newport psychiatric discharges 124 Newport psychiatric discharges 9.4 14.4 25.8 19.8 4.6 .69 8.1 .85* 11.1 .91 10.9 .92 * This reliability coefficient was based on a first half-second half rather than an odd-even correlation. Validity To test the validity of the PI it was ad- ministered to several new groups and scored by the previously devised stencil.1 These new groups were obtained at the U. S. Naval Train- ing Station, Newport, Rhode Island, the U. S. Naval Hospital, Chelsea, Massachusetts, and the U. S. Submarine Base, New London, Con- necticut. The group at Newport Naval Train- ing Station consisted of 508 inductees who had passed through psychiatric screening and were considered normal; 116 men who had RESTRICTED 40 ELIMINATING THE EMOTIONALLY UNFIT been under observation in the psychiatric ward but who were subsequently returned to their companies for trial duty; and 124 men who were being discharged for psychiatric reasons. The answer sheets for all three groups were It may be seen from Table 4 that cutoff scores can be selected which include most of the discharges and only a small percentage of the normals. For example, 3.9 per cent of the normals made scores of 18 or higher, while 51.6 per cent of the discharges made such scores. The difference between the percentages of dis- charges and normals identified by a cutting score of 18 was over 12 times its standard error.1 Similar analyses of the answer sheets from a smaller group of men tested at the U. S. Naval Hospital, Chelsea, Massachusetts, showed essentially the same ability to distin- guish between normal and psychiatric cases. A critical score of 18 identified 60 per cent of the psychiatric cases while including 3.9 per cent of the normals. Further evidence support- ing the value of the inventory came from analysis of the New London data. A comparison of average scores made by different Navy groups to whom the long form of the PI was given supplied additional infor- Table 4. Number and cumulative percentage at each score on Personal Inventory for 508 normals, 124 dis- charges, and 116 back-to-duty cases at Newport Naval Training Station. 508 Newport 124 Newport 116 Newport normals discharges back-to-duty Cumulative Cumulative Cumulative Score N per cent Score N per cent Score N per cent 0 1 0.19 0 0 0 0 1 5 1.17 1 0 1 2 1.72 2 21 5.30 2 2 1.61 2 4 5.17 3 17 8.64 3 0 1.61 3 5 9.48 4 34 15.35 4 2 3.22 4 6 14.66 5 33 21.82 5 3 5.64 5 8 21.56 6 32 28.11 6 5 9.68 6 11 31,06 7 40 36.08 7 2 11.29 7 10 39,68 8 38 43.56 8 4 14.52 8 15 52.58 9 47 52.81 9 5 18.56 9 10 61.20 10 51 62.84 10 6 23.40 10 11 70.70 11 40 70.71 11 9 30,65 11 5 75.01 12 35 77.69 12 1 31.46 12 10 83.63 13 22 81,92 13 9 38.71 13 2 85.35 14 25 86.84 14 2 40.32 14 1 86.21 15 19 90.58 15 4 43.55 15 4 89.66 16 17 93.82 16 4 46.78 16 0 89.66 17 11 96.08 17 2 48.39 17 3 92.25 18 7 97.45 18 4 51.62 18 0 92.25 19 3 98.04 19 2 53.23 19 2 93.97 20 3 98.63 20 4 56,46 20 2 95.69 21 0 98.63 21 3 58.88 21 1 96.55 22 3 99.22 22 2 60.49 22 0 96.55 23 1 99.41 23 6 65.33 23 1 97.41 24 1 99.60 24 5 69.37 24 1 98.27 25 0 99.60 25 3 71.79 25 0 98,2/ 26 1 99.89 26 1 72.60 26 1 99.13 27 0 99.89 27 4 75.83 27 0 99.13 28 0 99.89 28 0 75.83 28 0 99.13 29 0 99.89 29 4 79.06 29 1 100.00 30 0 99.89 30 1 79.87 31 0 99.89 31 2 81.48 32 0 99.89 32 5 85,52 33 0 99.89 33 0 85.52 34 1 100.00 34 2 87.13 35 0 87.13 36 2 88.74 37 6 93.58 38 1 94.39 39 1 95.20 40 1 96.01 41 2 97.62 42 1 98.43 43 0 98.43 44 0 98.43 45 1 99.20 52 1 100.00 Table 5. Personal Inventory scores for various Navy groups. Mean Personal Inv. Standard score deviation 276 Commissioned officers in submarine school 6.4 3.9 646 Coast Guard Reserve Cadets 7.4 4.6 189 Psychiatrically approved submarine school men 8.3 4.3 609 Newport normal recruits subsequently receiving ratings 8.3 4.2 1,512 Newport normal recruits 9.3 4.6 856 Newport normal recruits not receiving ratings within a year 10.0 4.8 444 USS' New Jersey men 11.1 5.7 74 Psychiatrically disap- proved submarine school men 14.6 8.4 2,301 Amphibious Forces men 15.2 8.0 301 Newport psychiatric dis- charges 21.6 11.0 mation on the validity of the Inventory. These data are presented in Table 5.6 The differentiating ability of the Inventory is shown graphically in Figure 1. The separate points through which the two curves are drawn scored by means of the previously devised stencil. The number and cumulative percentage of each of the three groups falling at each score is given in Table 4. RESTRICTED THE PERSONAL INVENTORY 41 —■ ■ ■ I 1 1 ■ , I I 1 BEST DIFFERENTIATION FOUND WITH PERSONAL INVENTORY, LONG FORM (REFERENCE 5) POOREST DIFFERENTIATION FOUND WITH PERSONAL INVENTORY, LONG FORM (REFERENCE II) Figure 1. Value of Personal Inventory, Long Form, in differentiating psychiatric discharges from normals. The two curves show the best and the worst differentiation found with the PI, Long Form. Each point, on each curve, represents a possible cutting score. Several such scores are indicated on each curve. The percentage of psychiatric discharges who equaled or exceeded each cutting score is shown on the vertical axis. The percentage of normals who equaled or exceeded each score is shown on the horizontal axis. Curve 1 is based on 508 normals and 188 discharges. Curve 2 is based on 307 normals and 89 discharges. represent different possible cutting scores. Each curve shows, along the horizontal axis, the percentage of a group of normal men who were included along with the percentage of psychiatric discharges indicated on the vertical axis. The two curves represent the best and the poorest differentiation found in any of the studies of the long form. Both are very much better than chance differentiation. In the best differentiation, 70 per cent of the psychiatric discharges were identified by a cutting score which included only 3 per cent of the normals. In the poorest differentiation, approximately 43 per cent of the discharges were identified by a cutting score which also included 3 per cent of the normals. RESTRICTED 42 ELIMINATING THE EMOTIONALLY UNFIT Expressing the results in terms of percent- ages gives an unduly favorable account of the usefulness of the inventory, for it neglects the fact that the satisfactory men far outnumber the unsatisfactory ones. At Newport Naval Training Station approximately 4 per cent of the men were given early psychiatric dis- charges. Thus about 40 were discharged of each 1,000 men examined. According to the best results shown in Table 1, approximately 28 of these 40 men could be identified by a cut- ting score which also picked up 29 of the 960 acceptable men in each 1,000. With less favor- able discrimination the same cutting score might include considerably more normal than psychiatric cases. However, administration of premium, as it usually was during the war, the PI can help materially in the selection process. 12 2 Enlisted Men’s Personal Inventory, Short Form' After considerable experience with the long form of the PI a short form of 20 items was constructed. This form includes the ten most discriminative items from the long form and ten other items which showed fairly high dis- crimination between normal and psychiatric groups. These were included to give continuity in the subject matter of the short form. The 20 Table 6. Personal Inventory, Short Form. In each question, mark (X) the answer which fits you one that fits you better than the other. best. Even if neither answer fits you very well, mark the 1 I graduated from high school Yes ( ) 1 No ( ) 2 I was a sickly child ( ) 2 I was an active child ( ) 3 I have felt bad more from head colds ( ) 3 I have felt bad more from dizziness ( ) 4 I seek excitement ( ) 4 I avoid excitement ( ) 5 I like to have people do things my way ( ) 5 I like to have people figure things out for me ( ) 6 I am more nervous ( ) 6 I am more easy going ( ) 7 Somehow I never could find enough to do in my free time ( ) 7 My free time always seemed to be filled ( ) 8 I wish I wouldn’t feel so tired ( ) 8 I wish I could have a more responsible job ( ) 9 I wish I could have more excitement ( ) 9 I wish I weren’t bothered by bad dreams ( ) 10 I wish I didn’t have so many aches and pains ( ) 10 I wish I wouldn’t keep changing my mind ( ) 11 I wish I weren’t so nervous ( ) 11 I wish I wouldn’t keep putting things off ( ) 12 I wish I could get myself to take more chances ( ) 12 I wish worrying wouldn’t make me sick to my stom- ach ( ) 13 I have more headaches than the average person Yes ( ) 13 No ( ) 14 The hours at night seem long Yes ( ) 14 No ( ) 15 I like most any kind of food ( ) 15 I have a poor appetite ( ) 16 After exertion I feel hungry ( ) 16 After exertion I feel dizzy ( ) 17 When excited I feel weak ( ) 17 When excited I feel stronger ( ) 18 I think I might like to watch a surgical operation sometime ( ) 18 The sight of blood upsets me ( ) 19 My heart sometimes speeds up for no reason at all Yes ( ) 19 No ( ) 20 I have never gone to a doctor for headaches or dizzy spells ( ) 20 I have occasionally gone to a doctor for headaches or dizzy spells ( ) the PI selects a small group (in the above case 57 men) which includes 70 per cent of all who would be discharged if the entire 1,000 were examined. Thus even though normal men were included, the amount of time necessary for psychiatric examinations could be greatly re- duced without overlooking a very large fraction of those men whom the psychiatrist would dis- charge if time permitted the examination of all recruits. If time of the psychiatrist is at a items are those marked with an (5) in the final column of Table 1. The short form is given in its entirety in Table 6. Scoring Key All items of the short form are scored. Table 7 shows the distribution of the critical ratios on two different populations. The correlation between the critical ratios determined in the two studies was .56. RESTRICTED THE PERSONAL INVENTORY 43 Table 7. Critical ratios of the items Inventory, short form. in the Personal Number Number Critical ratios (first study) (second study) 2.5- 4.9 1 5.0- 7.4 3 2 7.5- 9.9 7 4 10.0-12.4 6 7 12.5-14.9 1 5 15.0-17.4 3 1 20 20 Validity The short form was shown to differentiate between normal men and discharges as satis- factorily as the long form.9- 12 Curves similar to those shown in Figure 1 for the long form are presented in Figure 2 for the short form. The data pictured in the upper curve of Fig- ure 2 were obtained in a nonpredictive situa- tion. The 263 discharges had already been so classified and the 538 normals had been ac- cepted for duty. The data pictured in the lower curve were obtained in a predictive situation. The PI was administered and the answer sheets filed. Not until after the men had been accepted or discharged were the papers scored and the scores compared with the disposition of the men. Comparison of the two curves in Figure 2, or the data in Table 9, shows less sharp differ- entiation in the predictive than in the nonpre- dictive situation. For example, using a cutting score of 7, it was found that 60 per cent of the discharges could be predicted correctly in ad- vance of the psychiatric interview at a cost of including 9.6 per cent of the normals. Using the same cutting score in a nonpredictive situation, Table 8. Reliability coefficients (odd-even, Spear- man-Brown corrected) for the Personal Inventory, short form. Group rf 458 Newport recruits .66 536 Newport recruits .68 426 Amphibious Forces men .89* 257 Newport discharges .91 * This reliability coefficient was based on a first half—second half rather than an odd-even correlation. Reliability The short form proved to be as reliable as the long form. Table 8 shows the results from four groups of subjects. Table 9. Cumulative percentage falling at and above each and nonpredictive situations. score on the Personal Inventory, short form, predictive PI score 458 N Predictive use normals Cum. % N 30 discharges Cum. % N Nonpredictive use 538 normals 263 discharges Cum. % N Cum. % 20 0 0 0 0 19 1 0.22 0 0 5 1.90 18 0 0.22 1 3.33 0 14 7.22 17 0 0.22 0 3.33 1 0.19 15 12.93 16 0 0.22 0 3.33 0 0.19 17 19.39 15 1 0.44 0 3.33 0 0.19 23 28.14 14 1 0.66 2 10.00 2 0.56 18 34.98 13 0 0.66 1 13.33 0 0.56 20 42.59 12 0 0.66 0 13.33 1 0.74 10 46.39 11 4 1.53 5 30.00 4 1.49 15 52.09 10 6 2.84 3 40.00 7 2.79 24 61.22 9 4 3.71 3 50.00 2 3.16 12 65.78 8 9 5.68 1 53.33 7 4.46 8 68.82 7 18 9.61 2 60.00 11 6.51 8 71.86 6 22 14.41 0 60.00 37 13.38 11 76.05 5 42 23.58 3 70.00 42 21.19 14 81.37 4 55 35.59 1 73.33 63 32.90 7 84.03 3 84 53.93 4 86.67 92 50.00 9 87.45 2 94 75.11 1 90.00 124 73.05 18 94.30 1 93 94.76 2 96.67 106 92.75 11 98.48 0 24 100.00 1 100.00 39 100.00 4 100.00 RESTRICTED 44 ELIMINATING THE EMOTIONALLY UNFIT (T) NON* PREDICTIVE USE OF PERSONAL INVENTORY, SHORT FORM (REFERENCE 7) (g) PREDICTIVE USE OF PERSONAL INVENTORY, SHORT FORM (REFERENCE 9) Figure 2. Comparison of discriminative value of Personal Inventory, Short Form, in predictive and non- predictive use. This figure is to be interpreted in the same way as Figure 1. The curve for predictive use is based on 815 normals and 59 dis- charges: that for nonpredictive use on 538 normals and 263 discharges. Better differentiation was found in the nonpredictive than in the predictive situation. Compare with Figure 1 for relative discrimination value of long and short forms of PI. clearer differentiation between the two groups was achieved: 72 per cent of the discharges were identified at a cost of 6.4 per cent of the normals. But even the poorer differentiation shown in the predictive situation was as good as that shown in one of the nonpredictive studies using the long form (Figure 1). The usefulness of the short form in identifying an appreciable proportion of the psychiatrically undesirable men was unequivocally demon- strated. 4-2 3 Agreement between Long and Short Forms11 The relation between the long and short forms was shown by a correlation of .84 be- tween scores made on the two forms by a group RESTRICTED THE PERSONAL INVENTORY 45 of 426 men tested at the Amphibious Training Base, Solomons, Maryland. The correlation be- tween the long and short forms was essentially equal to the reliability of each. Furthermore, the two forms were equally valid. Choice between them became, therefore, largely a matter of preference. Both the long and the short forms continued in use throughout the war, with the Coast Guard and Maritime Service using the long form and the Navy adopting the short form (see Section 4.4). 4-2-4 Comparison with GCT in Identify- ing Psychiatric Cases11 The General Classification Test (GCT) was already given to all Navy recruits. It was, therefore, important to determine whether the PI worked more efficiently in selecting men scores varied from .01 to —.35. (The negative correlation was due to the fact that high scores on the GCT and low scores on the PI were con- sidered “good.”) The actual correlations for six groups are shown in Table 10. The PI was found to be more efficient than the GCT at any cutoff score in discriminating between normal men and those given psychiat- ric discharges. Table 11, for example, demon- strates that a score of 72 on the GCT identified 60.7 per cent of the psychiatric discharges but included 42.4 per cent of the normals. A score of 14 on the PI identified 59.6 per cent of the dis- charges and included 18.9 per cent of the nor- mals. Discrimination by the PI was obviously superior to that obtained by the GCT. The figures in Table 11 were obtained with the long form of the PI. No direct comparisons of the short form with the GCT are available. However, since the short form was found to correlate .84 with the long form, and since the two forms were shown to be approximately equal in validity and reliability, it seems highly probable that the short form, like the long one, is considerably more discriminative of psychiat- rically undesirable men than is the GCT. 4-2-5 Validation against Other Criteria All validation discussed above was obtained against the criterion, psychiatric judgment. It would be highly desirable to be able to present validation against criteria of combat profi- ciency or to use as criteria other records Table 10. Correlation between GCT and PI scores. Group r 1,071 Newport recruits — -.28 (old form GCT) 89 Newport psychiatric discharges .01 (old form GCT) 400 Submarine School candidates — -.17 (Otis test) 576 Amphibious Forces enlisted men — -.35 (new form GCT) 426 Amphibious Forces enlisted men — -.25 (new form GCT) 412 Newport recruits — -.28 (old form GCT) who would be given psychiatric discharges than did the GCT. The correlation between PI scores and GCT Table 11. Comparison of identification of discharges by GCT (old form) score and PI score. For normals, N = 307; for discharges, N = 89. GCT cutoff score Percentage discharges included Percentage normals included PI cutoff score Percentage discharges included Percentage normals included 34 10.1 0.7 37 10.1 0.0 43 20.2 1.3 29 20.2 0.0 51 30.3 6.2 24 31.5 0.3 59 38.2 14.3 20 41.6 2.3 67 50.6 28.7 17 49.4 6.2 72 60.7 42.4 14 59.6 18.9 75 70.8 49.8 12 68.5 30.3 81 79.8 66.8 10 79.8 49.8 87 91.0 86.0 7 88.8 74.6 97 100.0 99.4 2 100.0 99.7 RESTRICTED 46 ELIMINATING THE EMOTIONALLY UNFIT indicating how satisfactorily each man with- stood the pressures of military training and activity. Such criteria were extremely difficult to secure in any reliable form. An effort was, however, made to validate the PI against cri- teria other than the psychiatrist’s initial judg- ment. Comparison with Service Records10 One study compared PI scores with Service records one year after testing. The data are summarized in Table 12. The object of this study was to relate PI scores to the rating held one year after testing, to the conduct record The mean score for the discharges was 4.3 points higher than for the active men (critical ratio 3.9). Of the seven men who scored 26 or over on the PI, all were discharged during the first year of service. 2. The PI showed some tendency to differen- tiate between good and bad conduct cases. The mean score for the men with good conduct rec- ords was 1.7 points lower than for the men with some conduct offenses (critical ratio 4.1), and 2.8 points lower than for those with more serious offenses (critical ratio 4,6). 3. The PI showed some tendency to differen- tiate rated from nonrated men. The mean Table 12. Relation between Navy Service records and scores on the PI. Group N Mean a Sigma units of mean from general mean N Score of 18 and above % Entire group 1,466 9.3 4.6 63 4.3 Non rated men 856 10.0 4.8 +0.15 51 6.0 Rated men 609 8.3 4.2 -0.22 12 2.0 Nonoffenders 1,328 9.1 4.6 -0.04 52 3.9 Offenders 138 10.8 4.8 +0.33 11 8.0 Worst offenders 64 11.9 4.7 +0.57 7 10.9 Active 1,414 9.2 4.4 -0.02 52 3.7 Discharges 52 13.5 8.1 +0.91 11 21.2 Critical ratios of differences between Nonrated vs rated Nonoffenders vs offenders Nonoffenders vs worst offenders Active vs discharges means 6.1 4.1 4.6 3.9 during the first year of enlistment, and to the active or discharge status one year after test- ing. The records of 1,466 men were examined. All had been accepted by the psychiatrists at the time of original testing and boot training. Complete records were available on 1,007 of these men. Separate analyses of the total group of 1,466 and of the 1,007 men for whom com- plete records were available were in all re- spects practically identical. The principal re- sults of the analyses of the total group were as follows: 1. The PI identified a significant proportion of the 52 men who were later discharged. Twenty-one per cent of these men had received scores of 18 or above on the PI, as compared with but 4 per cent of those not discharged. score for the rated men was 1.7 points lower than for the nonrated men (critical ratio 6.1). 4. The GCT tended to differentiate dis- charges from those still on active duty, but to a decidedly lesser extent than did the PI, It tended to differentiate conduct cases to a slight degree. It differentiated rated men from un- rated ones somewhat more sharply than did the PI. This may have been due in part to the availability of the GCT scores in the assign- ment of ratings. While this evidence shows some relations be- tween PI scores and records one year later of conduct, rating, and discharge status, PI scores would have little value in selecting men for advancement. The restricted range of scores in this population should, however, be remem- RESTRICTED THE PERSONAL INVENTORY 47 bered in considering these data. All men in- cluded in the study had been approved by the examining psychiatrists. What results would have been obtained on an unselected population is unknown. Use of the PI in Selecting Paratroopers15 A second investigation using a criterion other than psychiatric judgment studied the usefulness of the PI in predicting success or failure in the parachute school at Fort Pen- ning, Georgia. Of 1,079 men tested there with the PI, 778 successfully completed parachute training. The remaining 301 men fell into 13 categories of failure. The four largest failure categories were: 89 men classified as having insufficient desire to continue training; 60 men who refused to jump from mock-up towers; 34 men who were permanently disqualified physi- cally; and 67 men who were injured in training and temporarily disqualified for varying lengths of time. Some of these would be expected to complete training later and others would later be classified as permanently disqualified phys- ically. The biserial correlation between PI score and completion or failure to complete para- chute training was —.39. The correlation for the Army General Classification Test was .26. (The population was a normal Army group in terms of Army General Classification Test scores.) These two correlations can be com- bined into a multiple correlation of .41 with success in parachute training. On the basis of these findings, it appeared that the PI could contribute to the improved selection of men for parachute training, and that the PI would be as effective for this purpose alone as in combi- nation with the Army General Classification Test. Escape Tank Training4 A third attempt to validate the PI against a performance criterion used success in escape tank training at the Submarine School, New London, Connecticut, as a criterion. The sub- marine escape training tank is a circular metal tank 100 feet in height filled with fresh water at approximately 90 F. Men reporting for training in the use of the submarine escape appliance, the “lung,” are first given a brief physical examination. They are then placed in a decompression chamber where they are in- structed in the principles on which the lung operates. After the pressure is reduced to normal, the men are given lungs and further instructions in their use. The training itself consists of going down a ladder into the water and using the lung. Later each man is taken down to a depth of 12 feet in a diving bell. He is required to ascend to the surface on an escape line twice. Finally he is required to make two satisfactory escapes from a lock at an 18-foot depth. Many men also escape from the 50- and 100-foot locks, but this is not re- quired. About 11/2 per cent of the men appeared to be seriously disturbed emotionally. This dis- turbance was shown by letting go the line and rising rapidly to the surface; spitting out the mouthpiece of the lung; coming up too fast; refusing to leave an escape chamber; excessive complaining or other signs of emotional dis- turbance. The PI was found to have a tetrachoric cor- relation of .14 with tank failure. In combina- tion with other tests, a multiple correlation of .43 appeared, but the PI alone could not be used satisfactorily to predict success or failure in tank escape training. Selecting Submarine Personnel18 The PI and a number of other tests were ad- ministered to all men processed at the New London Submarine Base in 1943 and the early part of 1944. These test scores were later com- pared with trait ratings of the men made by submarine officers. Scores on the PI, and on the other selection tests also, were found to be unrelated to the quality of performance aboard submarines as it was rated by officers. The absence of correla- tion can be attributed to the unreliability of the criterion and to the fact that the sample of men studied was already selected with respect to mental, physical, and emotional characteristics. Under these same conditions, the psychia- trist's evaluations of men judged to be emo- tionally unstable were no more useful in predicting the criterion than were scores on the PI. RESTRICTED 48 ELIMINATING THE EMOTIONALLY UNFIT Officers’ Personal Inventory Construction One form of the PI was constructed for use with officers. This form consists of 164 items. Forty-nine items were taken from the Long Form of the enlisted men’s PI with some changes in the wording of some items; the other 115 were new. In general format and in scoring method it was similar to the Long Form of the enlisted men’s PI. The officers’ PI is re- produced in the appendix. Scoring Key13 Arrangements were made for a validational study of the officers’ PI in the Amphibious Training Command, U. S. Atlantic Fleet. Two conditions made validation impossible. In eight months’ time only 83 officers were diagnosed as emotionally disqualified by the Boards of Medi- cal Examiners at the Amphibious Training Bases. Further, these Boards insisted on hav- ing access to the tentative PI scores at the time of their own examinations. The PI scores and the medical examiners’ ratings were therefore not independently arrived at. Consequently, the ratings could not be used as a basis for validat- ing the PI scores. The scores supplied to the Boards of Medical Examiners were based on a tentative 34-item scoring key. These 34 items are identical with ones contained in the scoring key of the long form of the PI used for enlisted men. Since the medical examiners did not have access to the individual answer sheets, they did not know how any man had answered the individual items. It was, therefore, possible to study the ability of each of the 164 items to discriminate between the 83 emotionally disqualified officers and a normal group of qualified officers. Thirteen hundred officers were included in the normal group. On the basis of item analysis comparisons between these two groups, a scoring key of 50 items was constructed. Each of the 50 items showed a critical ratio of 3.96 or greater in differentiating between the emotionally dis- qualified and the qualified officers of this study. A cutoff score of 30 on this key identified 53 per cent of the emotionally disqualified officers while including 2.6 per cent of the qualified officers. The odd-even reliability of this key is represented by a correlation coefficient of .88 corrected by the Spearman-Brown formula. Further study involving the application of this key to new groups of emotionally qualified and disqualified officers will be needed to establish its validity. In such further study the diagnos- ing medical officers must be without knowledge of the inventory scores at the time of diagnosis if validity of the PI for officers is to be estab- lished. Predicting Success in Marine Corps Officer Candidate School A study20 of the officers’ PI was conducted by the Medical Field Research Laboratory at Camp LeJeune, North Carolina, under better controlled conditions than were possible in the Amphibious Training Bases. At Camp LeJeune the officers’ PI was administered to 1,039 Marine Corps officer candidates. Each man was tested within one week after arrival at Camp LeJeune for a period of military training and observation for screening purposes. All men had V-12 college training and all had just com- pleted 12 weeks of recruit training at Parris Island. A follow-up study was made on the sub- jects after they had either completed or failed to complete officer candidate school four or five months later. Test results were seen by no one not connected with the research staff. A scoring key was constructed using the items which best differentiated between those who completed and those who failed to com- plete officer candidate school. Using this scor- ing key on the same group of men on whom it was constructed showed a biserial correlation of .48 between PI scores and success or failure in officer candidate school. When this scoring key was applied to new groups of 757 candi- dates21 and 671 candidates,22 correlations of .18 and .28 were found. For the two groups com- bined the correlation was .25. On these same groups the original NDRC-Navy scoring key correlated with the pass-fail criterion to the extent of .22 and .09. For the combined groups, the correlation was .16. These correlations in- dicate that the PI would not serve effectively as a single basis for selecting Marine Corps RESTRICTED OTHER METHODS OF SELECTING EMOTIONALLY UNSTABLE MEN 49 officer candidates. It may prove useful as part of a selection battery, however. Tentative data on this possibility were secured by the Medical Field Research Laboratory, Camp LeJeune, North Carolina, which found a multiple corre- lation of ,32 between the PI and a specially prepared confidential questionnaire and a pass- fail criterion in officer candidate school.23 4 3 OTHER METHODS OF SELECTING EMOTIONALLY UNSTABLE MEN 4'31 Tests Similar to the Personal Inventory New London-NDRC Questionnaire Reference 3 gives an analysis of the data obtained from testing 417 normal and 71 psychiatric ward cases at the Newport Naval Training Station by means of the New London- NDRC Questionnaire developed under an NDRC Division 7 project. This questionnaire did not appear to be as effective as the PI in distinguishing between acceptable and unac- ceptable men. Information concerning this questionnaire and its trial use at the U. S. Sub- marine Base, New London, Connecticut, is contained in reference 2. The Cornell Selectee Index The Cornell Selectee Index was developed under the Committee on Medical Research of OSRD. It used a different type of question (e.g., Did you ever regularly drink more than three quarts of whiskey a week?). A 32-item edition of the Cornell Selectee Index and the short form of the PI were tried out under iden- tical conditions on 1,000 recruits at Newport Naval Training Station. Both tests were taken before the recruits were given psychiatric ex- aminations, but the scores were not available until after psychiatric disposition of the men was complete. In this purely predictive situa- tion, the PI identified 61 per cent of the dis- charged men at a cost of falsely identifying 14 per cent of the normals. The Cornell Selectee Index picked up 67 per cent of the discharges and 19 per cent of the normals. Using the two tests together resulted in the correct identifi- cation by one or both of them of 82 per cent of the discharges and the false identification of 24 per cent of the normals. [Letter from Com- manding Officer, U. S. Naval Training School, Newport, Rhode Island, to Chief, Bureau of Medicine and Surgery, March 3, 1944, Serial No. 14939, P2-5 (1A)]. The Enlisted Personal Inventory, Form 2, NavPers 16845, adopted for general Navy use, contains items from the Cornell Selectee Index in addition to the 20 items of the short form of the PL The Personal Inventory, Army Air Forces Revision As part of the aviation psychology research program in the Army Air Forces, the Personal Inventory was revised for use in classification and redistribution of AAF officers and gunners. The history of the Aviation Psychology Pro- gram24 contains an account of the development of several modified forms of the PI and of the satisfactory use of these modifications. 4 3 2 Use of Battle Noise Equipment in Screening Psychiatrically Undesirable Personnel14, 11 Section 17.3 of NDRC developed, at Navy request, an extremely high-power, high-fidelity sound reproducing system. Upon completion, this equipment, the Battle Noise Equipment, was assigned to the U. S. Naval Training Sta- tion, Newport, Rhode Island. The medical offi- cers there were directed to study its potential usefulness in screening out those men who would be so disturbed emotionally by the noise of battle that they would be ineffective in com- bat. Assistance of the Applied Psychology Panel in conducting these studies was re- quested. Sound records of battle action were played at intensities of 110 db and higher to groups of men whose reactions to the noise were studied. During the playing of any sound sequence the men were observed by a psychia- trist who watched for any obvious signs of dis- turbance. In some cases the signs were very obvious. A few men broke ranks and ran away. RESTRICTED 50 ELIMINATING THE EMOTIONALLY UNFIT A few were taken directly to the psychiatric ward. Many more showed pallor, sweating, or other emotional signs. At the end of the noise exposure each man was asked to indicate, by holding up his hand or by writing on a card, which of a list of symptoms described by the experimenter he had experienced. Observations of this kind were made on nor- mal recruit groups and on men from the psychiatric ward. The sound was sometimes presented alone and sometimes as an accom- paniment to motion pictures of battle scenes. The motion pictures were shown to some men without noise. Finally, some men were not ex- posed to the pictures or the noise but were merely asked to check those symptoms they would feel if they heard the noises of battle. In each case the men were classified in accordance with the number of symptoms ex- hibited or reported. The resulting classification was then compared with the diagnosis arrived at in the regular psychiatric interview. It was found that the men’s reactions to the battle noise, either alone or with motion pic- tures of battle scenes, agreed fairly well with the results of the psychiatric interview. But it was also found that the men’s responses to the questions alone, without ever hearing the battle noises at all, served just as effectively in picking out those whom the psychiatrist con- sidered unfit for duty. Consequently no change in the psychiatric procedure was recommended and the battle noise equipment was not used for screening purposes. The Applied Psychology Panel recommended to the Navy that the equipment be used to give increased realism to some types of attack training. That recommendation was adopted and the equipment put to use in the Marine infiltration range at Camp Pendleton, California.16 SERVICE USE OF THE PERSONAL INVENTORY 4-41 Possible Uses The Personal Inventory might be used by the military Services in either of two ways: as a method of selecting men for special types of duty, or as a psychiatric screen for picking out those who should be given more detailed ex- amination. In view of the criteria which were available for validating the PI, there was little justification for using it as the basis for select- ing men for special duty without further study. There was much justification for using it as a psychiatric screen. In many installations the medical staff was not able to give thorough psychiatric inter- views to every man passing through the in- stallation. In such situations the PI provided a method of picking out a small group which contained most of the men who would have been discharged if time had permitted the care- ful examination of all men. Results obtained by using the short form of the PI in this fashion are presented in Table 13. This table is based on experience with a wartime recruit popula- tion in which about 5 per cent of the men were given psychiatric discharges. At other times, and with other groups, the absolute figures would differ throughout the table. But the prin- ciple of saving time by selecting for interview those men most likely to be rejected is appli- cable generally. Table 13 shows the saving in interview time which can be achieved by using the PI as a preliminary screen. For example, at a cutting score of 7, 156 men out of each 1,000 examined would be identified for interview. Of these 156 men, 32 would be discharged. These 32 consti- tute about 60 per cent of the total number of men who would be discharged had the whole 1,000 been examined. Thus 60 per cent of the potential discharges can be identified by a process which requires interviewing only 16 per cent of the entire population. If the psychi- atric examining facilities allow careful exami- nation of a larger fraction of the total flow of men, a lower cutting score can be adopted. For example, a score of 5 would require examining 308 of the 1,000 men. That group of 308 would contain 70 per cent of all who would be dis- charged if the whole 1,000 were examined. Using the information contained in Table 13, it is possible to adjust the cutting score in terms of the total number of men passing through a station and the time available for psychiatric examinations. RESTRICTED SERYICE USE OF THE PERSONAL INVENTORY 4 4 2 Psychiatric Endorsement Use of the Personal Inventory was endorsed by the Subcommittee on Psychiatry of the Na- tional Research Council’s Division of Medical Sciences at a meeting on June 29, 1944, in a resolution which contained the following rec- ommendations : “NOW THEREFORE BE IT RECOM- Ground Forces, and the Marine Corps. One or more forms were adopted for routine screening purposes by the U. S. Navy, Coast Guard, and Maritime Service. A special form was developed by the Army Air Forces21 (see Section 4.3.1). Format A (the long form with a scoring system giving separate scores on each of 20 clusters of items) was adopted as part of the routine selection procedure at the U. S. Sub- Table 13. Efficiency of the Personal Inventory in reducing the number of men to be interviewed. Cutting score Number identified for interview per 1,000 Number interviewees to be discharged Number not interviewed, who would be discharged had they been interviewed 20 0 0 53 19 2 0 53 18 4 2 51 17 4 2 51 16 4 2 51 15 5 2 51 14 12 5 48 13 16 7 46 12 19 7 46 11 39 16 37 10 56 21 32 9 79 26 27 8 110 28 25 7 156 32 21 6 212 32 21 5 308 37 16 4 422 39 14 3 594 46 7 2 786 47 6 1 956 51 2 0 1,000 53 0 MENDED that these methods be put to use as widely as possible in the military services, and that further studies along these lines be continued and extended with a view to the progressive improvement and refinement of these tests in the light of validating data from actual performance under conditions of military service and combat; BE IT FURTHER RECOMMENDED that tests of a similar nature be developed for use in the selection of personnel for all types of special assignment.” Service Adoptions of the PI The PI was used experimentally in a number of places by the Navy, Coast Guard, U. S. Maritime Service, Army Air Forces, Army marine Base, New London, Connecticut. On May 22, 1944, a directive from the Bureau of Naval Personnel specified a score of 14 or lower on Format B for admission to Submarine School. Format B (the long form with a simpler scoring procedure giving only one score) was adopted as part of the routine selection pro- cedure at the U. S. Navy Intercept Officer Training School at St. Simons Island, Georgia; at the U. S. Maritime Service Training Sta- tions located at Sheepshead Bay, New York; Catalina Island, California; and St. Peters- burg, Florida; and in the U. S. Coast Guard classification centers. On January 24, 1945, the Director of Enlisted Personnel wrote to the commanders and com- manding officers of some 50 receiving stations, RESTRICTED 52 ELIMINATING THE EMOTIONALLY UNFIT naval training centers, and training schools in- forming them that the short form of the PI (Enlisted Personal Inventory, Form 2, NavPers 16845) “has been adopted by this Bu- reau to aid in identifying enlisted personnel who lack the emotional stability required by submarine and amphibious duty.” “Men mak- ing scores of 8 or higher . , . are ineligible for transfer to Submarine School, New London.” Instructions were given to make careful psy- chiatric examinations of all men in the Am- phibious Forces making scores of 10 or above. The Officers’ Personal Inventory was not adopted for routine use in any Service. 45 RECOMMENDATIONS FOR FURTHER WORK The PI worked more successfully in military selection than it will in selecting civilians for peacetime jobs. On all personality tests of this type it is possible for a man to make a higher or a lower score than he deserves, for the nor- mal, nonsymptomatic answers to the questions are sometimes fairly obvious. The most prob- able civilian use of the PI would be in selecting men for positions for which the men them- selves apply. Under these conditions there will be considerable temptation to give “normal” answers to the questions and thus to make a better score than is deserved. In wartime the motivation is to a large ex- tent reversed. Some men answer the questions in such a way as to make a poorer or less nor- mal score than they otherwise would, hoping thereby to avoid military duty and particu- larly the more hazardous types of duty. The PI, consequently, lumps together those who are really emotionally unstable and those who are willing to admit more psychiatric symptoms than they normally exhibit. Since both of these groups should be examined in detail by the psychiatrists, this grouping is not a serious fault in military use of the PL The distinction between wartime and peace- time conditions of use of this test is pointed out here to warn against any expectation that the success of the PI in screening out men unsuitable for military duty can be duplicated in peacetime industrial situations. How well it will work in peacetime military situations re- mains to be seen. Efforts to construct personality tests which are completely free from the possibility of fak- ing have met with little success. Nevertheless these efforts should be continued. Tests of emo- tional stability and personality traits could be used with much greater confidence if the men being tested were unable to secure better scores than they deserve by giving false answers to the test items. The PI was originally intended as a means of predicting a man’s ability to perform satis- factorily in hazardous duty or in situations re- quiring great emotional stability. In practice it became a device for predicting what a psychia- trist would say about a man on the basis of a short interview. A thoroughgoing validation of the psychiatrist’s own judgment is needed. Fu- ture work on instruments similar to the PI should be accompanied by parallel studies of the psychiatrists’ judgment. Both of these methods of predicting a man’s emotional sta- bility should be validated against an independ- ent criterion, preferably one based on combat performance. The difficulty of securing such criteria is great, but that difficulty must be overcome if instruments similar to the PI are to develop beyond the stage of being predictors of the unvalidated judgment of the psychia- trist. RESTRICTED Chapter 5 DETERMINING VOCATIONAL INTERESTS Dael Wolfle a Summary An inventory of interests was constructed _ for possible use in the classification of military personnel. The inventory was in- tended to secure information on a wide variety of interests, e.g., vocational, social, recrea- tional, scientific, and religious. Information of this type could provide a valuable supplement to that obtained by interviews, vocational his- tory, or standarized tests of ability, in classify- ing men for military duty. The interest inventory, a scoring key, and detailed statistical information were turned over to the War Department Adjutant Gen- eral’s Office. The inventory has some ad- vantages over other existing interest ques- tionnaires. It has not yet been validated for military personnel. 51 INTRODUCTION Information regarding a man’s interests combined with information about his ability ought to provide a better basis for job assign- ment than would either type of information alone. Previous attempts to secure information about interests in standardized and comparable form were not considered satisfactory for mili- tary use. The Adjutant General’s Office there- fore requested the Applied Psychology Panel to construct an interest inventory for use in the classification of military personnel. The work was assigned, as Project SOS-7, to a contract with Harvard University. 32 THE ACTIVITY-PREFERENCE TEST Construction of Items Several paper-and-pencil tests of interest were already in existence at the time this project work began. The two most widely used were those developed by E. K. Strong and G. F. Kuder. Both attempted to determine a man’s interests by asking a number of questions about the activities which he liked or disliked. Typical of the Strong Vocational Interest Blank for Men was a list of related items (such as hobbies or school subjects). The sub- ject marked each to show whether he liked, dis- liked, or was indifferent to that activity (hobby, school subject, etc.). The Kuder Pref- erence Record consisted of paired items, or groups of three items. The subject indicated the most and least preferred activity within each group. The Activity-Preference Test differed from both earlier tests in several respects, although it continued the tradition of attempting to de- termine general types and fields of interests by asking a series of questions on quite specific likes and dislikes. It differed in attempting to cover a wider range of activities; in attempting to make the items compared more directly com- petitive with each other in terms of time, cost, and opportunity than was true of the earlier tests; and in including questions on biographi- cal background and attitudes toward oneself. The Activity-Preference Test contained a total of 220 exercises. Each exercise, except for items of factual information included in Sec- tion I, consisted of four alternative statements, each of which described an activity or situa- tion. Taking the test required selecting the most liked and the least liked one of the four activities described, or selecting the most likely and the least likely behavior or response under the conditions stated. The item form was, therefore, similar to that used in the Kuder Preference Record. Two examples are: Upon a free afternoon, which of the following would you like MOST and which would you like LEAST? Exercise 74: A. to go to a vaudeville show B. to help organize a Boy Scout Troop C. to take part in a play D. to take the family out for a picnic Think of some situation in the past few years in a This chapter is based entirely on the final report of Project SOS-7.1 RESTRICTED 53 54 DETERMINING VOCATIONAL INTERESTS which you have succeeded markedly. Which of these do you think was MOST and which do you think was LEAST responsible for the success? Exercise 22: A. I always worked hard to earn quick money. B. I had an excellent memory for people’s faces. C. I was not timid in making constructive sugges- tions. D. I put more energy than most people into getting a job done. The 220 exercises were divided into seven sections. Section I (10 exercises) covered biographi- cal information concerning the person when he was ISVa years old. Sections II and V (15 exercises each) cov- ered preferences and interests as they existed at age 151/2. Sections III and VI (75 exercises each) cov- ered preferences at the time of taking the in- ventory. Sections IV and VII (15 items each) re- quired the subject to look ahead to the time when he would be 48 years old and to indicate the activities which he expected to prefer at that age. Two forms of the inventory contained the same exercises, but in different orders. The like most-like least form required two responses to every exercise and thereby avoided the systematic error consequent to a rosy or blue outlook which in other interest tests has resulted in excessive marking of things as being “liked” or as being “not liked.” There may be a “general liking of things” function, and in some connections it may be important. The present instrument had evi- dence upon this only in the first, or biographi- cal, section. Even though one person “likes most everything” more than a second, still the lesser level of intensity of feeling of the second person is the intensity level that dictates his conduct. His contrasting interests at the low level may well be just as important in deter- mining his vocational preference as the more intense level of the first person. A choice must be made, no matter at what level. The four options within each exercise were very carefully devised so that for the average person there is no moral issue, or right or wrong, or good or bad choice, involved. Any expression of interest can be falsified, but if the subject sees no reason for dissimulation, and if there is no earmarked “right” or “wrong” answer, an honest expression of con- viction can be expected. It is true that an in- telligent person can produce an answer sheet which is biased in a direction which has been chosen by the person if he is so inclined. Such bias is not as great as might be expected be- cause of the subtlety of the scoring schemes. The internal evidence suggested that a useful falsification—useful in the sense desired by the subject—of the activity-preference record is measurably less with this instrument than with any of the other commonly available interest tests. This difficulty of falsification was accom- plished by the form employed and by the care directed to this issue in the construction of exercises. As a general interest measure serving all subjects, the exercises were devised to sample as nearly as possible all fields of life in which preferences can be shown. A careful scrutiny was made of the waking activities of the entire day and of the seasons, spring, summer, fall, and winter; of the social contacts of life from the narrowest social group, the home, to the widest; of the earliest age (chosen as age 1S1 involving the maturity and initial out- looks of adulthood, and of the last age presum- ably thought of as still in full prime (chosen as age 48). The results of these surveys were reflected in the wide variety of exercises in- cluded in the inventory. It was also set down as a requirement that the four options constituting an exercise should be genuinely competitive. A subject was not asked to express a preference between a trip to Mexico, requiring much time and money, and seeing a football game, or to ex- press a preference between swimming and skating, or between other activities which in the nature of things do not simultaneously claim attention. 5 2-2 Types of Interests Covered The 220 exercises were planned to secure in- formation on 40 different types of interests (rubrics). These were: RESTRICTED THE ACTIVITY-PREFERENCE TEST 55 Rubric No. Description 1 Things and mechanisms (not a repetitive or routine use). 2 Spatial thinking (visual-mindedness). 3 Calculation, computation, number facility (not symbolic thinking). 4 Orderliness, system, organization, regimen- tation. 5 Routine, undisturbing activities vs desire for change, variety, and suspense. 6 Monetary motivation and personally gain- ful activity. 7 Masculinity-femininity. 8 Family and neighborhood. 9 Adventure, daring, risk vs self-security, timidity. 10 Power (the leadership involving or based upon control) vs submission and satis- faction in position of menial servitude. 11 Gross physical activity in work and/or play. 12 Fine dextrous activity and craftsmanship. 13 Gregarious—bold and forward vs shy. 14 Handling people for their own presumed good. 15 Independent, self-sufficient vs gregarious. 16 Memory activity. 17 Pioneering and initiative (leadership based thereon) vs traditionalism. 18 Instructing, advising, supervising. 19 Salesmanship (not over-the-counter). 20 Verbal expression—oral and written. 21 Vocational level from professional to labor. 22 Verbal reception: reading and hearing. 23 Competitiveness. 24 Humor. 25 Music: participation and appreciation. 26 Arts: graphic, plastic, artistic: participa- tion and appreciation. 27 Abstract vs concrete thinking. 28 Experimental verification vs fiat and a priori conviction. 29 Principal and suspended judgment vs im- pulse, instinct, sensuousness. 30 Practical reality vs fantasy. 31 Acting. 32 Religion and idealism. 33 Self-confidence vs tendency to worry. 34 Emotionality vs stolidity. 35 Perseverance and industry vs laziness and shillyshallying. 36 Outdoor vs indoor work, play, and amuse- ment. 37 Plants and animals. 38 Integrated behavior vs neuroticism and compensatory adjustments. 39 Aesthetic and synaesthetic responsiveness. 40 Vehement, explosive, violent vs restrained, calm, quiet. An a priori set of scoring weights was made up which permitted the tabulation of scores on each of these 40 rubrics. Each rubric score was obtained by totaling the weights assigned to each answer of each exercise considered appro- priate for that rubric. First Administration The Adjutant General’s Office arranged to administer the Activity-Preference Test to groups of enlisted men at three Army training centers. A total of 3,734 men was tested. Papers from a subgroup of 2,201 men all born in the years 1918 to 1925 were used for subse- quent statistical analysis. ° 2 4 Item Study Reliabilities of the rubric scores were com- puted using a priori weights for the data from Sections III and VI (total of 150 exercises at present age). The reliabilities of the separate scores ranged from .00 to .79. Eliminating the least reliable rubrics reduced their number to 34. The range of uncorrected reliabilities for the retained rubrics was .26 to .79. Refined sets of scoring weights were obtained for 15 rubrics by analyzing the responses to each exercise scored for each of these 15 ru- brics. The reliabilities of the scores based on refined weights for the exercises in Sections III and VI ranged from .50 to .84. Estimated rubric reliabilities for the entire inventory ranged from .55 to ,89, with a median of .75. 5-20 Factor Analysis The variances for the 34 retained rubrics were adjusted in terms of the judged impor- tance of each. The covariances were modified accordingly. The matrix of these adjusted vari- ances and covariances was analyzed by the method of principal components. The matrix was analyzed into 34 components, the first five of which accounted for 72.31 per cent of the total variance. The full meaning of these components de- pends upon their relations to success in a variety of occupations. Any formal naming of them is RESTRICTED 56 DETERMINING VOCATIONAL INTERESTS difficult and likely to be misleading until a large amount of knowledge of their vocational rela- tions and usefulness accumulates. The desig- nations listed below were intended only for purposes of identification. Following each, the component variance is given, and then an indi- cation of the predominant rubrics or personal characteristics important in each component. MIMSEC, the first component (36.41) M Masculine I Isolationists (prefers being alone) M Mechanical S Social E Effeminate C Conversational PEPGAP, the second component (14.60) P Persevering and pioneering E Economic P Practical and interest in power G Gross (rough or boisterous, as opposed to refined) A Adventurous and daring P Physical activity (interest in) PAMRIM, the third component (10.37) P Power and gross physical activity (interest in) A Aggressive (salesmanship and competition) M Money (interest in) R Religious I Industrious M Music, mechanisms, and craftsmanship (interest in) RODPOD, the fourth component (5.71) R Routine and religion (interest in) O Orderly D Domestic P Pioneering (interest in) O Outdoors (interest in) D Daring NEVCOM\ the fifth component (5.23) N Nature-loving E Y | EVangelistic—religion with salesmanship C ) Interest in general COMpetence, that is, positive O f weightings of power, vocational level, mechanisms, M f spatial, orderliness, verbal reception and expression, ’ and music. The names given to the components were each of two syllables which are equally well pronounced together if reversed, the first re- lated to the rubrics having positive weightings in the component and the second to those having negative weightings. A person scoring high on the first component may be called a mimsec, and one scoring low a secmim person. The letters mim are related to rubrics with positive weight- ings, and the letters sec to rubrics with negative weightings. To those initiated into the nature and uses of this preference instrument, these letters may provide reasonable but very tenta- tive pegs around which to build a more com- plete picture, and to the uninitiated (in general the person taking the test for the first time) these words will be harmless in that none of them suggest a type of “expected” or “right” response. 5-2-6 Preparation of Scoring Key A final scoring key for each of the five prin- cipal components was constructed. Each item in Sections III and VI was assigned a weight within the range of —4 to +4, and each item in Sections I, II, IV, V, and VII a weight within the range of —6 to +6. For any one of the five components, most of the item weights were zero. The reliabilities of the five component scores are shown in Table 1. These reliabilities were determined by correlating scores on one half of the indicated section or sections with scores on the other half of the indicated section or sections and then stepping up by the Spearman- Brown formula. Table 1. Reliabilities of component scores.* (N = 200). Sections of test II & III & IV & Total Components I V VI VII test I MI MSEC .77 .87 .86 .79 .94 II PEPGAP .54 .67 .83 . 56 .86 III PAMRIM .80 . 76 .87 .51 .91 IV RODPOD .83 .83 .78 .53 .89 V NEVCOM .76 .61 .61 .55 .77 * It was estimated that with refined weightings for the items in Sections III and VI, the reliabilities of the several components for the entire instru- ment would be .96 for Component I; .89 for Component II; .93 for Com- ponent III; .89 for Component IV; and .81 for Component V. The reliabilities are high enough to justify separating the inventory, for some purposes, into two equivalent tests, each half as long as the original. The estimated reliabilities of a half-length test for each component score would be: MIMSEC = .89 PEPGAP = .75 PAMRIM = .84 RODPOD = .80 NEVCOM = .63 RESTRICTED THE ACTIVITY-PREFERENCE TEST 57 Final Disposition The test, a table of scoring weights for the five components, and detailed statistical infor- mation concerning the test and the factor anal- ysis were submitted, by request, to the Office of the Adjutant General at the end of 1944. That office had agreed to be responsible for validating the test and for conducting such additional sta- tistical analysis of the data as seemed desirable and useful. The pressure of other duties prevented the Personnel Research Section, AGO, from com- pleting this work. It remained, at the end of the war, in the form in which the Applied Psy- chology Panel turned it over to the Army. The individual items are promising in that they have no easily picked out “right” answers. But considerable further work will eventually be required to determine how and under what conditions it can be best used to aid in the classification of Service personnel. RESTRICTED Chapter 6 THE SELECTION OF RADIO CODE OPERATORS By Dael Wolflea Summary At navy request the Applied Psychology _ Panel developed an improved test of radio code aptitude. The test first requires the subject to learn three simple characters of the radio code and then measures his ability to differen- tiate these characters as they are sent at faster and faster speeds. The test was demonstrated to have a validity of .50 and to be superior to all other available code aptitude tests. It was officially adopted by both Army and Navy for use in the selection of men for radio code training. 61 INTRODUCTION The process of acquiring skill in the recep- tion of International Morse Code can be broken down into two phases. In the first phase the student learns 36 characters representing the alphabet and the digits. This task is relatively simple and ordinarily requires only a few hours or at most a few days of practice. In the second phase the student must learn to respond cor- rectly to the characters when they are sent more and more rapidly. Ordinarily from 10 to 20 weeks of practice are required before the student acquires enough speed and skill to be- come a useful operator. Some men fail to learn the individual char- acters satisfactorily, and many who do learn them fail to acquire sufficient speed to be use- ful as military or commercial operators. The result is a high attrition rate unless the stu- dents are very carefully preselected for code learning ability. Failure rates as high as 40 to 50 per cent were found in some schools early in World War II. The problem of selecting code operators be- came apparent in World War I, and efforts were made at that time to develop suitable tests for that purpose. One of the tests developed then became standardized as the Signal Corps Code Aptitude Test (SCCAT) early in the nineteen-twenties. It remained as the standard selection test until after World War II began. Its low validity made the need for a better test obvious. When the Committee on Service Personnel— Selection and Training was organized in June 1942, one of its early requests was to establish a research project on the selection and training of radio code operators for the Navy. The di- rective specifically ordered that research should be conducted to improve the techniques of se- lecting personnel for assignment to radio code instruction. The request was accepted and assigned to the Psychological Corporation, as the contractor, as NDRC Project N-107. At- tempts to build better code tests were being made at the same time by other groups, both military and civilian. 6.2 TYPES OF CODE APTITUDE TESTS Three chief types of code aptitude tests have been tried. 1. Discrimination tests measure the ability of the subject to distinguish between complex rhythmic patterns of dots and dashes. SCCAT is the best known example. 2. Code learning tests measure the ability of the subject to learn code characters. The test developed by L. L. Thurstone and an Army test (ROA-2) are examples. 3. Speed of response tests measure the ability to copy a few easily learned code characters at rapid rates of transmission. The NDRC Speed of Response (SOR) test is the best example. The NDRC project worked with all three types. Preliminary results showed the third to be most promising, so work was concentrated on the development of a speed of response test. a This chapter is based primarily upon the work of NDRC Project N-107. 58 RESTRICTED THE NDRC SPEED OF RESPONSE TEST OF CODE APTITUDE 59 6 3 THE NDRC SPEED OF RESPONSE TEST OF CODE APTITUDE4 6-31 Development The idea for this type of test came from Biegel.6 After introducing his students to the radio code characters, he presented signals to them in groups of 30, with less and less space between the individual characters in the suc- cessive groups. His test was long, apparently consisting of 900 characters in addition to the introductory learning series. Further, it was administered three times in order to increase its reliability. His data indicated that it ap- peared to have a reasonably high value in pre- dicting subsequent achievement in radio code classes. A speed of response type of test has two parts. First, there is a learning unit in which the subjects are taught a few characters. There have been several opinions as to the degree of difficulty of the letters to be taught. NDRC Project N-107 finally chose three very easy characters on the principle that it was desir- able to have nearly all the subjects able to copy the characters at a slow speed of presentation, so that the test itself would be almost entirely of the ability to respond quickly to these signals. Second, there is a testing unit in which the sub- jects are tested on these learned characters at increasingly faster rates of presentation. In the first edition the letters D S (...) and R (. . .) were used. After these three letters were learned, the test itself began. It consisted of the three letters, D, S, and R, sent in random order, first at slow speeds and then more and more rapidly until the fastest speed was reached. It was then repeated in re- verse order going from the faster down to the slower speeds, with the slowest speed omitted. This initial form was found to have consid- erable predictive ability, but it contained a number of faults. The test was too long. The initial part, however, was too short; some of the students failed to learn the three characters well enough to distinguish them properly even at the slowest speed. Form 2 attempted to correct these faults by using three easier letters, I (. .), N (-.), and T (-), by increasing the amount of learning drill, and by shortening the test itself. The test went through two other preliminary editions in which refinements were made and in which it was adapted to machine scoring. Final editions were prepared, one for the Army and one for the Navy, a manual of instructions was written,3 and the SOR was put into routine use for selecting radio operators for military duty. 6 3 2 Validity Several criteria can be used to determine the validity of a code aptitude test. The final grade assigned in code school usually depends upon several factors, only one of which is code re- ceiving speed. Even so it has been used in some studies, as has the dichotomous criterion of failing or passing the course. Code receiving speed as determined by school examinations has also been used. The chief difficulty with this criterion is that both test content and admin- Table 1 . Validity of Speed of Response Test. Study conducted by Criterion used Weeks of instruction No. of classes No. of men Median validity coefficient NDRC CRT 5 2 216 .30 NDRC CRT 8 10 1,932 .34 Army CRT 7-8 4 178 .41 Army Code speed 7-8 4 221 . 55 Navy Code speed 14 1 185 .37 Navy Pass-fail* 14 1 299 .49 Thurstone Code speed* 6 1 215 ,61 Thurstone Pass-fail* 16 1 196 . 57 *Biserial correlation; others were Pearson correlations RESTRICTED 60 THE SELECTION OF RADIO CODE OPERATORS istration procedures vary from school to school. In order to provide a standardized criterion, Project N-107 developed the Code Receiving Test (CRT).1’2 This test consisted of phono- amounts of code instruction. Eight of the re- sulting validity coefficients are reproduced in Table 1. They range from .30 to .61, with half of them .49 or higher. The validity shown by these correlations was considerably higher than that of the Signal Corps Code Aptitude Test which had been the standard selection test for approximately 20 years. Validity of the SOR is shown in terms of practical usefulness in predicting code school success in Figures 1 and 2. Figure 1 analyzes the records of 1,001 men after eight weeks of training. The men in the class who had achieved speeds of 14 groups per minute or better, those who had achieved 12 groups per minute or better, those who had not yet gotten better than 8 groups per minute, and those who had HIGH SCORES ON SOR INTERMEDIATE SCORES ON S 0 R LOW SCORES ON SOR Figure 1. Relation between scores on Speed of Response Test and performance in radio code school. Men were first classified as high (10 to 15 per cent of sample), intermediate (10 to 15 per cent), or low (70 to 80 per cent) in terms of score on SOR. The vertical bars show the percentage of each group having code speeds, after eight weeks of training, shown at the bottom of the figure. For example, 12 per cent of the high-score group, but only 3 per cent of the low-score group, had achieved speeds of 14 groups per minute or higher. N = 1,001. graphically recorded instructions, practice ex- ercises, and test content. The test content con- sisted of five-character groups in which the 36 characters each appeared with equal frequency. Letters and numbers were presented in a purely random order, so they frequently occurred in the same group, CRT was therefore more diffi- cult than the usual school test of receiving speed; a speed of 14 groups per minute being the equivalent of 20 words per minute of plain language material. The entire set of CRT rec- ords included tests at 12 speeds varying from 4 to 25 groups per minute. The validity of SOR was determined by sev- eral investigators, using different criteria, on different groups of subjects, and after different Figure 2. Speed of Response Test scores and code speed after 12 weeks of training. The distribution of code speeds attained in 12 weeks of training is shown for three groups of men classified accord- ing to their SOR scores and for the total group of 616 men. Low—bottom 70 to 80 per cent in SOR scores Int.—next 10 to 15 per cent in SOR scores High—top 10 to 15 per cent in SOR scores All—the three groups combined already been eliminated from code school are shown, divided according to whether their SOR test scores were high, intermediate, or low. Validity of the test is shown by the fact that RESTRICTED THE NDRC SPEED OF RESPONSE TEST OF CODE APTITUDE 61 men who made high SOR scores more frequently attained high speeds, and were less frequently found in the poor groups, than men whose SOR scores were intermediate or low. In Figure 2, results are shown in a different way for 616 men after 12 weeks of code train- ing. Again, the men making high SOR scores less frequently failed and more frequently at- tained higher speeds than did men making low or intermediate SOR scores. Comparison with Other Tests The evidence summarized in Section 6.3.2 demonstrated the validity of the SOR. But other code tests were also available, and the Army and Navy wanted to adopt the most sat- isfactory one. Each Service, therefore, con- ducted an experimental comparison of several possible tests. The Army comparison8 of six tests, using two criteria, was based on a total of 221 men in four classes at Scott Field. The results are summarized in Table 2. The SOR was found to have a slightly higher validity than the Thurstone7 test. Both were considerably better than any of the Army radio code tests or the Army General Classification Test in selecting good operators. were all considerably less efficient than the Thurstone and NDRC code aptitude tests. Table 3. Navy comparison of six tests for predict- ing success in radio code training. Test Correlation with final aver- 14th week age course Pass-fail code speed grade in course SOR .37 .28 .49 Thurstone’s test .36 .31 .42 Clerical aptitude .06 .14 .26 Spelling .21 .22 .28 GCT .12 .15 .09 Reading .19 .18 .16 The superiority of SOR over Thurstone’s7 test, as shown in Tables 2 and 3, was slight. It was, however, a shorter test. The playing time of the records was 27 minutes for SOR and 44 minutes for Thurstone’s test. Both tests had the necessary characteristics of being simple to administer, suitable for group administra- tion, and designed for scoring with the Inter- national Business Machines Company scoring machine. 6.3.4 Adoption by Army and Navy The Army and Navy tests8-9 referred to in Section 6.3.3 both led to official adoption in 1944 of the SOR for routine Service use. The Coast Guard and Maritime Service also used it. Special recordings of the test were made by the project for Army and Navy use. The Navy edition was recorded on two sides of one record to be played at 33Vs rpm. It is identified as Radio Code Test: Speed of Re- sponse, Form 2, Columbia Recording Corpora- tion serial number YTNY 2902-2903. The an- swer sheet is NavPers 16573. The Army edition was recorded on eight sides (four records) to be played at 78 rpm. The test content is identical with the Navy form; minor differences in instructions were made to adapt it to Army use. The Army renamed the test Army Radio Code Aptitude Test, 1944, ARC-1. The records bear Columbia Recording Corpora- tion serial numbers from XP33522 to XP33529. The answer sheet is IBM Form I.T.S. 1100 B 1133. Table 2. Army comparison of six tests for predict- ing success in radio code training. (N = 221) Test Correlation with School speed CRT speed SOR .55 .41 Thurstone’s test .47 .44 ROA, X-l .33 .39 SCCAT, I .30 SCCAT, II .33 AGCT .24 The Navy comparison9 of six tests, using three criteria, was based on samples of 183 to 306 men in the radioman school at Bainbridge Naval Training Center. The results are sum- marized in Table 3. The SOR and Thurstone tests were again found to have the highest validity coefficients. The other four tests (Cler- ical Aptitude, Spelling, Reading, and GCT), RESTRICTED 62 THE SELECTION OF RADIO CODE OPERATORS 6.3.5 Advantages for Military Classification The Navy data reported in Table 3 showed low correlations of code proficiency with the General Classification Test and tests of spelling, reading, and clerical ability. Age was also shown to have a low correlation with achieve- ment in code school; for nine separate classes of from 123 to 258 men each, the median cor- relation with age was —.14 (range —.22 to +.03). Similarly, low correlations were found between code proficiency and education, me- chanical ability tests, and tests of subject mat- ter knowledge.4 Since youth, intelligence, education, and me- chanical understanding are at a great premium for many military assignments, it is highly advantageous to have a test which predicts success in code school satisfactorily but does not correlate highly with these other variables. This situation permits both a negative and a positive type of selection for code training. Men with superior qualifications for other assignments can be eliminated from consider- ation, and selection then made from the re- mainder. By this means competent radio oper- ators can be secured without taking men who are especially desirable for other assign- ments. 6 4 SELECTION FOR OSCILLOSCOPE CODE RECEPTION In 1944 the Naval Research Laboratory re- quested Applied Psychology Panel assistance on problems of training men to receive code signals presented visually on an oscilloscope screen. In the course of training a number of men to receive such code, information was se- cured on the relation between SOR scores and learning records.5 In a group of 57 men a bi- serial correlation of .50 was found with a cri- terion of meeting or failing to meet a specified proficiency level after three weeks of training. Rank order coefficients of .57 and .52 were found between SOR scores and proficiency at two code speeds. These correlations are as high as those obtained with auditory code. A visual adaptation of the SOR should be tried out, even though it might not work better than the present auditory test. Present results indicate that the principle of the SOR test may be as useful in selecting men for visual code reception as for auditory code. RESTRICTED Chapter 7 SELECTING RADAR OPERATORS Donald B. Lindsleya Summary A number of tests were designed for use in selecting and screening radar operators. The tests were intended to measure aptitude for the visual tasks of reading and interpreting oscilloscope patterns. Most of the tests showed satisfactory reliability. Validity data were un- satisfactory because of the lack of an adequate criterion of proficiency for radar operators. Two of the tests were adopted by the Navy for use in the Combat Information Center Aptitude Test, Form 2. This test showed cor- relations of .45, .55, and .56 with final grades in three successive classes at naval training school (tactical radar). 71 INTRODUCTION Despite the fact that radar played a critical part in helping win World War II and was from the start recognized as an important and spe- cialized operation, little was done to establish standards of operating proficiency and to select operators in terms of such standards. The rapid development, production, and extension of field uses of radar meant that thousands of men had to be trained quickly. Training schools mush- roomed almost overnight. These schools were usually overcrowded and understaffed; fre- quently there was insufficient equipment for proper teaching and demonstration; and in order to meet quotas, training was often ex- tremely brief. Although a selection problem existed continually, few selection standards were developed and these were seldom rigidly applied due to the continued demand for oper- ators. This demand prevailed throughout World War II since new types of radar meant new trainees or retraining of previously trained operators. Toward the end, when the demand for operators abated slightly, the number of men being inducted was smaller and the quality poorer; consequently the number available for training was reduced and it was still difficult to apply rigorous selection standards. The principal criteria of Army and Navy specifications for radar operators were as fol- lows : age, under thirty; intelligence, slightly above average; vision, average or corrected to average; and interest or experience in related fields. The establishment of these criteria was arbitrary. Early in 1943 the Army requested a research project on psychological factors in radar oper- ation. A Navy request for similar studies was made shortly afterward. The work was as- signed, as Project SC-70, NS-146, to the Yerkes Laboratory of Primate Biology (Yale Univer- sity). One of the first tasks undertaken was to construct a series of tests for the purpose of selecting and screening radar operators. Re- quests from Army and Navy training centers called for screening measures to eliminate men already assigned for radar training. It was hoped that such screening would result in meet- ing quotas with more proficient operators and also save valuable training time and facilities otherwise wasted on men who were unable to pass the course and meet minimal standards of proficiency. Not only were selection standards poor, but in many instances little attention was paid to final proficiency; the inept were often passed and took their places as relatively ineffi- cient operators. After a survey of the types of functions per- formed by radar operators, it was concluded that one of the critical aspects of the radar operator’s task was visual perception, espe- cially the ability to detect visual changes quickly and accurately, and to note their move- ment, form, and spatial relationships. Sustained attention and emotional stability under pres- sure also were recognized as important factors. A series of pencil-and-paper tests, believed to embody measurements of some of these ca- a This chapter is based upon the work of Project SC-70, NS-146. RESTRICTED 63 64 SELECTING RADAR OPERATORS pacities, was constructed. These tests were then tried out in Army and Navy training cen- ters in order to determine their reliability and validity. Although reliability was readily ascer- tained, validity presented a problem which was not satisfactorily solved. Validity required a suitable criterion of proficiency against which to correlate the test results. The principal cri- teria of proficiency available at the time were school grades and instructor ratings of per- formance during training. Neither of these proved to be reliable measures. The examina- tions upon which course grades were based fre- quently were weighted heavily with nonessen- tials and seldom were comparable in difficulty from class to class or from school to school. Ratings of performance presented similar diffi- culties. Later it was possible to demonstrate that objective ratings of performance could be devised and that proficiency measures could be developed which were objective and discrimina- tive. In retrospect it now appears that it would have been profitable to have concentrated ini- tially on the development of proficiency meas- ures to be used at the termination of training, in order that a satisfactory criterion of per- formance might have been available for corre- lation with test results. 72 DEVELOPMENT OF SELECTION TESTS FOR RADAR OPERATORS In May 1943, Project SC-70, NS-146, pre- pared a series of thirteen pencil-and-paper Oscilloscope Operator Tests.1’2 The tests meas- ured perceptual and speed capacities thought to be essential to radar operation. Table 1 names the tests and summarizes the chief data avail- able concerning each. The tests are printed in reference 2, Some of the tests, especially Nos. 4, 5, 9, 10, 11, 12, and 13, simulated closely some of the tasks which a radar operator is required to perform and therefore had apparent “face validity.” For example, the Ratio Estima- tion Test presented a series of items which required the estimation of the ratio of paired vertical lines of different height. This was a duplication of the task presented by the A-scope on a ground control interception [GCI] radar used for determining elevation of targets, where the ratio of two simultaneous pips had to be judged by the operator. The Coordinate Read- ing Test presented a simulated plan position indicator [PPI] scope face on which there were a large number of blips whose range and bear- ing were to be read. Similarly the other tests reproduced some aspects of the operator’s task. Thus quantitative measures of speed and ac- curacy of performance could be obtained. 7-2-1 Reliabilities and Intercorrelations of the Tests Drew Field Study After preliminary tryout with enlisted men at Camp Murphy, Florida, 12 of the 13 tests were administered to 378 radar operators-in- training at Drew Field. The purpose of this study3 was to determine the reliability and validity of the tests as selection or screening in- struments. Unfortunately no satisfactory vali- dation data could be obtained in this study since the criterion of operator proficiency which con- sisted of school grades and ratings proved to be unreliable statistically and therefore inadequate for correlation with the test results. Distribu- tions of scores for most of the tests were found to give adequate discrimination. The reliability and intercorrelations of the tests, based on a sample of 100 randomly selected men, were de- termined. These data are presented in Table 1. The reliability of all tests was satisfactory, except for the scale reading and oscilloscope reading tests, which were subsequently revised and lengthened. Except in a few instances suffi- cient independence of measurement was found to warrant further experimental study of the individual tests with the view to establishment of predictive batteries. The correlations of the tests with the Army General Classification Test (AGCT) were all positive and ranged from .19 to .63. It was recommended that the AGCT be included in any battery of tests for the selection of radar operators. The tests had adequate reliability for indi- vidual prediction. Although intercorrelations indicated that the tests dealt with related func- RESTRICTED DEVELOPMENT OF SELECTION TESTS FOR RADAR OPERATORS 65 tions, it was decided that two or more of them might well be included in the same selection battery. It was further decided to try the tests out in another situation in which more adequate validating data might be obtained. Accordingly, the following study was made. Orlando Study An adequate criterion of radar operator per- formance was believed to be available in the eight operating stations of the Army Air Forces Tactical Center, Orlando, Florida.7 Here men of varying degrees of experience in operational training were performing operating duties daily Most of the tests gave promising validity coefficients; the six providing the best predic- tion possibilities were as follows: Scale Read- ing, Course Location, Plot Reading, Ratio Estimation, Coordinate Reading, and Polar- Grid Coordinate. The validity coefficients for the first four of these tests, in the individual stations, ranged from .35 to .78 with 17 of the 19 coefficients above .50 and 9 above .60. Co- ordinate Reading, measuring proficiency in PPI reading, and Ratio Estimation, measuring abil- ity to estimate pip ratios and to match pips, were shown to distinguish between different levels of training and experience. The Plot Table 1. Information concerning Oscilloscope Operator Tests.2,3,7> 9 Name of, test Face validity Reliability (split-half, corrected) Mean correla- tion with other tests Correlation with AGCT Mean validity RHO7 Used by Navyf 1. Form Detection .96 .33 .22 2. Form Conversion I .97 .43 .39 3. Form Conversion II .92 .37 .45 4. Scale Reading X .81 .45 .63 .60 CIC 5. Oscilloscope Reading X .86 .40 .30 6. Course Location .74 7. Spot Location .98 .42 .20 8. Target-Course Analysis .96 .27 .19 9. Plot Reading X .94 .42 .42 .65 10. Ratio Estimation X .96 .39 .47 * TR 11. Coordinate Reading X .90 .47 .45 * TR 12. Coordinate Plotting X .94 .46 .43 13. Polar-Grid Coordinate X .90 .51 .47 .53 J V /CIC ♦Discriminated significantly among operators with different amounts of training. fTR = Tactical Radar Aptitude Test (NavPers 16574). CIC=CIC (Combat Information Center) Aptitude Test, Form 2 (NavPers 16980). under the supervision of experienced instruc- tors. Two officers in charge at each station rated the men on operating performance. Ratings were made on A-scope reading, plotting ability, PPI scope reading and GCI operation. The reliabilities of the ratings in five of the eight stations were found to be satisfactory. Accord- ingly, a study of the relationship between the test results and the ratings of operator profi- ciency in these five stations was made. The different stations served Early Warning (EW), Chain Home Link (CHL) and Ground Control Interception (GCI) functions. Correlations were computed for the separate stations; the number of men in individual stations ranged from 14 to 36. Reading, Scale Reading, and Polar-Grid Co- ordinate tests were particularly applicable as selection devices for predicting success in vari- ous types of radar operation. Other Results Two of the above tests, Polar-Grid Coordi- nate and Scale Reading, were given at the Naval Training School, Virginia Beach, Virginia, to 120 search radar operators. The principal vali- dating criterion was the final course grade. The correlations between test results and final grades in the radar course were .50 and .55 respectively. During the summer of 1943, three additional tests were constructed. These were the Airborne RESTRICTED 66 SELECTING RADAR OPERATORS Oscilloscope Reading Test, Oscilloscope Conver- sion Test, and Oscilloscope Interpretation Test. The first two are concerned with airborne radar operator functions, especially interpretation and transmission of symbolic directions from operator to pilot. Both tests were tried out on AI (airborne interception) operators at Boca Raton Field and were found to be not highly predictive of success as measured by school grades. The Oscilloscope Interpretation Test is a test of ability to detect different types of radar signals through varying degrees of back- ground masking or jamming. This test was administered at the Naval Training School, Fort Lauderdale, Florida. Reliability was sat- isfactory, but validation data could not be ob- tained. The test was tried out by the Services in an AAF navigation school for selecting radar operator trainees and also in a special research project in the Eighth Air Force. 7 2 2 Application of the Tests by the Services Army In May 1944, upon request from the Psycho- logical Research Section of the Air Surgeon’s Office, copies of the following tests were sup- plied for experimental tryout in selecting path- finder crews in the Eighth Air Force: Target- Course Analysis, Coordinate Plotting, Form Conversion I, Form Conversion II, Oscilloscope Reading, and Airborne Oscilloscope Reading tests. The results of these studies are not known. In the summer of 1944, upon request from the Air Surgeon’s Office, special machine-scored editions of the Scale Reading Test and the Oscilloscope Interpretation Test were made available for use at Selman Field, Louisiana, in the selection of radar operator trainees. The Polar-Grid Test was revised by representatives of this station for use in connection with the same problem. The Coordinate Reading Test and the Oscil- loscope Interpretation Test were tried out in a battery of tests for the selection of bom- bardier instructors at Midland Field, Texas. It is understood that the tests gave sufficient promise to be incorporated in a radar operator selection battery. Navy Tactical Radar Aptitude Test.8’9 Three of the tests, the Polar-Grid Coordinate Test, Ratio Estimation Test, and Coordinate Reading Test were included in the Tactical Radar Aptitude Test (NavPers 16574) which was included as part of the selection requirements established for naval training school (tactical radar). The Polar-Grid Coordinate Test measures the examinee’s ability to translate the reading of a point on a polar coordinate to a grid co- ordinate chart. The Ratio Estimation Test measures the ability to estimate the relative lengths of lines presented in pairs. The Co- ordinate Reading Test measures the ability to estimate the direction and range of targets on a polar coordinate chart. The Bureau of Naval Personnel tried out the Tactical Radar Aptitude Test on several classes of approximately 100 men each.9 Using several criteria of success in Tactical Radar School, validity coefficients were generally small, rang- ing from —.19 to +.34. Twenty-seven of a total of 40 correlations were .20 or lower. The CIC Aptitude Test. On the basis of these results, however, the Bureau of Naval Person- nel constructed a new test, the CIC Aptitude Test, Form 2 (NavPers 16980). This test retained the Polar-Grid Coordinate Test from the Tactical Radar Aptitude Test but dropped the other two. In their places were in- cluded the Scale Reading Test and a Relative Movement Test which was originally developed by the University of California Division of War Research under Section 6.1 of NDRC. At the same time that the selection tests were being revised and improved, work was under way in improving the examinations and final grades given at NTSch (tactical radar). As a result of both of these improvements, the CIC Aptitude Test Form 2 gave much higher, and much more consistent, prediction of school suc- cess than had the Tactical Radar Aptitude Test. Detailed results are shown in Table 2. Correla- tions with final grades in three successive classes were .45, .55, and .56. These results RESTRICTED PROFICIENCY MEASURES AND RADAR TRAINERS AS SCREENING DEVICES 67 demonstrated the usefulness of the test as a selection device. A minimum Navy standard score of 50 was, therefore, included as one of the selection requirements to be observed in the selection of candidates for tactical radar train- ing. 73 RELATIONS BETWEEN VISUAL PROFICIENCY AND RADAR OPERATION The following study was made to determine whether there was a relationship between cer- tain visual capacities as measured by the Bausch and Lomb Ortho-Rater and proficiency in radar operation. This study0 was carried out in field operating stations of the Army Air Forces School of Applied Tactics, Orlando, Florida. Visual measurements, including binocular and monocular visual acuity (near and far), verti- 74 USE OF PROFICIENCY MEASURES AND RADAR TRAINERS AS SCREENING DEVICES The usual procedure is for men to proceed from “boot camp” or a classification center to a training center for primary training in an assigned specialty. In the case of radar, there- fore, it would be highly desirable to have selec- tion standards which could be applied before a man is assigned to training, in order to elim- inate those who would probably be “washed out” in training. There is, however, another point at which screening and selection may occur. This is after primary, but before advanced or operational, training. Also it frequently happens that small groups are selected from larger groups for specialized duty or performance which calls for exceptional skill and ability. For these sec- ondary types of selection the use of proficiency Table 2. Reliability and validity of CIC Aptitude Test, Form 2. Reliability Correlations with CIC achievement test Class 7L Class 8L Class 9L Correlations with final grades Class 7L Class 8L Class 9L Polar-Grid Coordinate .85 .49 .37 .42 .39 .45 .43 Scale Reading .85 . 55 38 .56 .46 .51 .56 Relative Movement .82 .48 40 .47 .31 .51 .39 Total test .92 .61 .45 .57 .45 .55 .56 cal and lateral muscle balance (phoria), stere- opsis, and color vision, were obtained for 157 radar operators. Ratings of proficiency in scope operation were also obtained. In general it was found that operators with substandard binoc- idar acuity at near distances were rated as less proficient than those with normal or better acuity; also that operators with excessive over- convergence at near distances were rated lower than those with normal convergence. In view of these results, tests of binocular acuity (near) and lateral phoria (near) were recommended for examination of prospective radar operators. It was recommended that minimum standards be set at 1.0 (decimal notation) for the near acuity test and at 6 prism diopters of overcon- vergence or esophoria. or achievement measures of an objective and performance type is recommended. Radar pro- ficiency measures and their use in the classi- fication and assignment of men are described in references 4 and 5. Such measures, particu- larly where knowledge of radar equipment and its functional uses was involved, have been shown to differentiate between expert, average, and apprentice levels of performance. Thus the proficiency measure may be used at the end of one stage of training to select those who will advance to the next stage. Frequently a man will do well in courses where knowledge of equipment is required but will fail in operational performance. In so far as certain radar trainers provide good simula- tion of operating conditions and at the same RESTRICTED 68 SELECTING RADAR OPERATORS time make possible quantitative measures of performance, which are often impossible to se- cure with the actual equipment, they are val- uable for assessing some of the performance aspects of radar operation. The measures ob- tained may be used to eliminate men who would not profit from further training. This implies, of course, that a significant relationship exists between performance on the trainer and oper- ational performance with the actual gear. Dur- ing the war there were a few radar trainers which seemed to meet requirements for use as selection devices. It should be emphasized also that, if the relationship between trainer per- formance and actual radar operation is high enough, the trainer may be used as a validating criterion against which to correlate the results of pencil-and-paper radar aptitude tests. This is particularly important since one of the chief obstacles in the development of radar aptitude tests has been in finding suitable and adequate criteria of proficiency of radar operation. Finally, if the validation of radar aptitude tests must rest on success or proficiency at- tained in various stages of training rather than upon final operational proficiency, which may be impractical to measure, it should be empha- sized that effort should be placed initially upon the improvement of proficiency or achievement measures. These must be made comprehensive and objective; performance functions should be emphasized and should be given a weighting proportional to their value in the final assess- ment of the operator. Without a reliable and adequate measure of proficiency it is impossible to establish the validity of a selection test. 75 FUTURE DEVELOPMENT OF RADAR APTITUDE TESTS There are a variety of radar applications for land, sea, and air operations, each type of equip- ment differing in its uses, characteristics, and operating procedures. It is probable that oper- ator requirements differ for each application and that selection standards will have to vary accordingly. During the course of World War II, minimal use of radar aptitude selection devices occurred and comparatively little prog- ress was made in developing and introducing such selection measures. The main pitfall seemed to be the problem of finding an ade- quate validating criterion. The solution to this problem lies in working out a plan for utilizing operational criteria, either from overall field performance or from specialized, objective methods of evaluating operational performance during training. The improvement and objecti- fication of proficiency measures, such as achievement examinations, ratings of perform- ance, or radar trainer scores, is a necessary prerequisite to the development of better selec- tion tests. RESTRICTED Chapter 8 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTFINDER OPERATORS By William E. Kappauf, Jr.a Summary Early in its history NDRC undertook work on the improvement of stereoscopic heightfinders. Efforts to improve this equip- ment led quickly to the development of related efforts to improve the selection and training of heightfinder operators. This psychological work was directed at first by Division 7. When a specialized psychological section was estab- lished in June 1942 (first as the National Re- search Council Committee on Service Personnel —Selection and Training, and later as the Ap- plied Psychology Panel, NDRC), responsibility for the work was transferred to it. Standards for the selection of stereoscopic heightfinder operators were developed, recom- mended to and adopted by the Army, and vali- dated by the project. Assistance was given to the Army in setting up stereoscopic testing centers where men to be trained as heightfinder operators were selected. When the Navy established a school for train- ing men for the rate of fire controlman (R), the Army heightfinder standards were used as the basis for setting selection requirements for rangefinder operators. The Applied Psychology Panel work was extended at the same time to include studies of the selection and training of rangefinder operators. Improved selection standards were recommended to the Navy and adopted for general use. The project staff trained Navy personnel for duty in the two testing centers where men to be trained as rangefinder operators were selected. In the course of the work on improving selec- tion standards, information was obtained on the reliability of a number of visual test in- struments : the Shuron pupillometer, the NDRC interpupillometer, the modified Massachusetts vision test kit, the Ortho-Rater, the projection eikonometer, the stereoscopic trainer M2, the vectograph-pursuit test, and the Dearborn- Johnston test. Correlations among the several tests of stere- oscopic vision were low enough to indicate that these tests were not measuring the same kind of ability and could not be used as substitutes for each other. A thorough study of military visual requirements and of improved methods of visual testing is recommended. 81 PROBLEMS IN VISUAL SELECTION The obvious importance of good vision in a wide variety of military situations and tasks led to a great deal of work on visual tests and visual standards. The Applied Psychology Panel was active in part of this work, dealing in par- ticular with problems of stereoscopic vision and night vision. Very little is known regarding the actual visual requirements of most jobs, or the mini- mum visual standards which should be imposed in selecting men for a particular job. The prac- tice, in view of this ignorance, has often been to set very high visual requirements for any military job in which vision appears to be im- portant. When this is done for many jobs, it soon becomes difficult to find enough men who meet the standards. The result is that require- ments are then completely disregarded or are modified by classification personnel who are usually unfamiliar with the relative importance of the several standards. To be reasonable, selection standards should set the minimum re- quirements which will keep school or training failures down to a tolerable number but not a This chapter is based on work done by the Princeton University Fire Control Research Project of NDRC Division 7, and by Project N-114 and the Heightfinder Project of the Applied Psychology Panel. A supple- mentary report is contained in Volume 2 of the Sum- mary Technical Report of Division 7. RESTRICTED 69 70 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTF1NDER OPERATORS impose impossible requirements on a classifi- cation department. The Applied Psychology Panel was asked to investigate the severity of the selection stand- ards for stereoscopic rangefinder and height- finder operators. The work led to the develop- ment and improvement of some of the testing equipment, to validation experiments, and to some developments in instrument maintenance and operator training to ensure dependable validation test results. 82 HISTORICAL SUMMARY OF WORK ON RANGEFINDER AND HEIGHT- FINDER SELECTION TESTS Qualifications for stereoscopic heightfinder operators were listed in the Army heightfinder pamphlet in use before the war. These quali- fications included 20/20 vision or better in each eye, hyperphoria less than V2 prism diopter, esophoria less than 12 prism diopters or exo- phoria less than 6 prism diopters, interpupil- lary distance between 58 and 72 mm, general good health, and no history of eyestrain or fatigue. These qualifications had been estab- lished on an a priori, not an empirical, basis. Their application rested entirely with the bat- tery commanders at the time that men were to be detached for training at the heightfinder school. Some of the men who arrived at the school met the standards; others did not, as indicated by visual acuity and other tests given them at the school. Interviews revealed that many men assigned to the school had no interest in heightfinder work. Selection was certainly not the best, and this coupled with lack of stu- dent interest meant mediocre heightfinder op- eration. 821 Studies at Fort Monroe, Virginia When the Princeton University Fire Control Research Project was set up under Division 7 early in 1941, one of its objectives was to for- mulate a satisfactory selection program for heightfinder operators. The new tests were to be tests of demonstrated merit. Accordingly, a wide variety of visual, physiological, and pencil- and-paper tests were given to each class of students at the school, and the relation between these test scores and final school grades was determined. By the end of the year, a good number of the original tests had been elim- inated from further consideration in the test battery. Validity was indicated for certain tests of stereoscopic vision (notably three tests on an instrument called the projection eikonometer and one on the stereoscopic trainer M2) and for pencil-and-paper tests of intelligence and mechanical aptitude.27 The task of obtaining satisfactory validation data had been difficult because the total number of students who had passed through the heightfinder school was less than 150 and they, through such selection as was applied, were more homogeneous than the general run of enlisted personnel. Finally in February 1942, Division 7 proposed a tentative new selection battery,1 Suggested minimum qualifications for stereo- scopic heightfinder or rangefinder operators were: General intelligence—an Army standard score’3 of at least 100 on the AGCT, or at least 85 on the Navy O’Rourke GOT. (These were erroneously believed to be equivalent scores.) Mechanical comprehension—a grade of I, II, or III on Part 3 of the Army test MA2 or MAS, or at least 75 on the Navy O’Rourke Mechanical Com- prehension Test. Height—for Army only—not less than 5 feet, 6 inches (imposed by tripod height for Army heightfinders). Inter pupillary distance—60 to 70 mm (imposed by limits of instrument adjustment). Vision—visual acuity of at least 20/20 in each eye— hyperphoria not greater than V2 prism diopter —exophoria not greater than 6 prism diopters— esophoria not greater than 6 prism diopters. Stereoscopic vision—an Army standard score of at least 110 (V20 above mean) on the projection eikonometer, and a standard score of at least 110 on the stereoscopic trainer M2 modified to include power driven change of range. Desire for training—selecting only those men who indicated that they wanted training as a stereo- scopic observer after the nature of an observer’s duties had been explained. b Army standard scores have a mean of 100 and a standard deviation of 20. RESTRICTED HISTORICAL SUMMARY OF WORK ON SELECTION TESTS 71 It was further proposed that special personnel be trained to administer these tests at selection centers. The results of the testing for each man were to be recorded on his qualification card. With the cooperation of the Coast Artillery Corps, the proposed tests were administered for standardization purposes at Fort Eustis, Virginia. Analysis of the records from these standardization experiments indicated where improved administration was required and where test simplification was possible. In par- ticular it was found that the projection eiko- nometer test could be reduced from three dif- ferent tests to one, the stereo-vertical test.c So that the Services could extend the pro- gram, detailed descriptions of the test equip- ment and testing procedures were provided.2-3-G The equipment included the Shuron pupil- lometer for measuring interpupillary distance, a modified Massachusetts vision test kit for measuring visual acuity and phoria, a projec- tion eikonometer as developed from an instru- ment used earlier at the Dartmouth Eye Institute (Section 8.3.7), and a modified stere- oscopic trainer M2. In accordance with the original proposals of Division 7, three stere- oscopic testing centers were established, at Fort Eustis, Camp Wallace, and Camp Callan. A complete summary of the data which pro- vided the basis for the Division 7 selection recommendations is found in reference 27. It is of interest to note what tests showed no useful or significant relation to heightfinder performance. They included a static test of stereoscopic vision, a series of pencil-and-paper tests other than those concerned with general intelligence and mechanical comprehension, and a wide variety of ocular and general physio- logical tests. Visual acuity and phoria require- ments were retained in the selection battery specifically for their usefulness in reducing testing time by weeding out men who would have difficulty passing the stereoscopic tests. Validation data on acuity and phoria were meager because the heightfinder school students did not show significant departure from normal in these respects. Studies at Camp Davis, North Carolina There were two specific and recognized limi- tations in the conditions of the experiments which lay behind the Division 7 recommenda- tions for the selection of stereoscopic operators. The men tested in the school and on whom the tests were validated were too homogeneous. The criterion score which was used as the measure of heightfinder operation was a vari- ability score, not a score of absolute accuracy. Since only 4 per cent of entering selectees could pass all the established selection requirements, there was speculation as to the possibility of getting comparably efficient selection with lower selection cost by using other stereoscopic vision standards, A follow-up validation study was therefore requested by the Antiaircraft Artillery Command. The second validation experiment was con- ducted by the Applied Psychology Panel project on the Selection and Training of Heightfinder Operators.12 This project, accepting the ade- quacy of existing selection requirements other than the stereoscopic ones, proposed to deter- mine the most satisfactory selection test(s) of stereoscopic vision. Through the cooperation of the Antiaircraft Artillery Command, a series of classes with men of all levels of stereoscopic ability were admitted to the Heightfinder School at Camp Davis. Through the further cooperation of the A.A. Schools at Camp Davis, and the Heightfinder School staff in particular, the school heightfinders were so maintained and reference measures of target position were so determined that it was possible to score the students not only on the variability of their performance but also on their absolute accu- racy. Four tests of stereoscopic vision were administered to all students prior to enrollment at the school. Two of these were the projection c Throughout this STR and in the original research literature this test is called the projection eikonometer test. The name, however, confuses the test with others used by the Dartmouth Eye Institute in the analysis of a visual condition known as aniseikonia. To avoid this confusion it is strongly recommended that in future Service use, the present test be called the stereo-vertical test. RESTRICTED 72 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTFINDER OPERATORS eikonometer test and the stereoscopic trainer test which Division 7 had recommended. The others were tests which had been developed under a Division 7 laboratory contract with Harvard University. All students were men who had expressed an interest in learning to operate stereoscopic instruments. The data of the validation study, based on a validation population of 92, indicated that the most efficient selection test was the projection eikonometer (see data in Section 8.4.4). Fur- thermore, if the passing score on this test was raised from 110 to 115, the stereoscopic trainer test could be eliminated from the test battery. Use of the projection eikonometer test with a passing score of 115 eliminated fewer men in the selection process and admitted a school pop- lation of which a greater portion could be ex- pected to graduate. The practical results of the research were these: a change in the stereoscopic vision test requirement in the selection battery for height- finder operators, a small increase in the number of selectees who could pass all tests (from about 4 per cent to about per cent), and a decrease in the amount of testing time required in order to qualify or disqualify a man. 8 2 3 Studies at Fort Lauderdale, Florida On the basis of the foregoing research, the Navy set up a list of selection requirements for men entering schools to be trained for the rate of fire controlman (R). This was a new rate established in June 1943, and was to be awarded to men proficient in the operation of rangefinder and radar equipment. The selection standards adopted at that time were as follows: General intelligence—a Navy standard scorecl of 55 on the GCT. Mechanical aptitude—a Navy standard score of 45 on the MAT. Reading ability—a Navy standard score of 50 on the Reading Test. Interpupillary distance—between 60 and 70 mm. Vision—visual acuity 20/20 in each eye uncorrected. —hyperphoria not greater than Vz prism diopter. —exophoria not greater than 6 prism diopters. —esophoria not greater than 6 prism diopters. Personal qualifications—emotionally stable, calmness under stress, quick reactions, not over 30 years of age. As a result of recommendations from the Ap- plied Psychology Panel to the Bureau of Naval Personnel and the Bureau of Medicine and Surgery, these requirements were modified in April 1944 in three respects. General intelligence—the passing score on the GCT was dropped to 50. Vision—visual acuity of 20/22 or better was accept- able for acuity measurements made on the Ortho- Rater (test raw score of 9). Stereoscopic vision—the requirement added was a Navy standard score of at least 58 (cr = 10) on the multiple projection eikonometer. Having adopted these selection standards, the Navy set up two selection centers, one at Sampson and the other at Farragut, to classify men for training for fire controlman (R). The two testing centers were staffed by men trained in the administration of the selection test bat- tery by members of Project N-114, the Applied Psychology Panel. At the further request of the Navy, Project N-114 carried out an experimental investigation of the adequacy of this set of selection stand- ards for the Fire Control School at Fort Lauder- dale, Florida. Actually this research program began before the above selection requirements were in effect, and so the student groups which were tested and examined at the school were representative of the unselected Navy popula- tion. The project sought a test battery which would be efficient in selecting men who could pass the theory and study part of the course and also succeed in becoming good rangefinder operators. Compiled course grades on theory and operation and a variability score for range- finding during the final examination were taken as the validating criteria. The criterion popula- tion for the rangefinder operator tests num- bered 141 men. The result of this investigation22 was the development of two test batteries which were found to be equally effective in terms of picking good students. Battery I below required longer testing time per man but passed more men than did Battery II. The batteries proposed were: d Navy standard scores have a mean of 50 and a standard deviation of 10. RESTRICTED DESCRIPTION AND RELIABILITY OF TESTS USED 73 Selection Selection Battery I Battery II General intelligence GCT 50 GCT 45 Arithmetic ability AR 45 AR 55 Interpupillary distance 60-70 mm 60-70 mm Visual acuity and phoria Visual acuity score of at least 9 for each eye on the Ortho- Rater Hyperphoria less than V2 prism diopter (Ortho-Rater score 4 to 7) Esophoria less than SV2 prism diopters, or exophoria less than AV2 prism diopters (Ortho-Rater score 4 to 12) Stereoscopic vision Multiple proj. Ortho-Rater eikonometer depth score score of at of at least 5. least 58. Personal qualifications As in the original Navy re- quirements ; must want to become rangefinder operator. Johnston test. Navy pencil-and-paper tests were also investigated to determine their use- fulness in selecting men for fire control courses. 8-3-1 The g}luron Pupillometer This instrument for measuring interpupil- lary distance was used in all the early selection work. It was first adopted and recommended for selection use by Division 7.2> 3 It consists of a frame with two sliding glass panels. On each piece of glass is a vertical hairline. When the subject’s eyes assume parallel lines of re- gard and the examiner bisects each of the sub- ject’s pupils with one of the hairlines, the dis- tance between the hairlines is the measured interpupillary distance. The median of three measurements with this instrument has an average error of somewhat less than 0.5 mm. 8-3 2 The NDRC Interpupillometer The NDRC interpupillometer17 was devel- oped to meet the need for accurate interpupil- lary distance adjustment in rangefinder and heightfinder operation. It is described in detail in Chapter 22 of Volume 2, Applied Psychology Panel. It incorporates the principle of the Shuron pupillometer in an instrument where the subject makes his own slide settings. The median of three measurements taken with this instrument has an average error of less than 0.25 mm.14 The NDRC interpupillometer was used in the Navy selection program. These instruments were not available in quantity for use in train- ing stations and aboard ship, so each man had his interpupillary distance accurately measured at the selection center. The measurement was recorded on his qualifications card, for use when and if he began training for the job of rangefinder operator. 8-3-3 Interpupillary Distance Measures in Selection Selection on the basis of interpupillary dis- tance is necessary because the range of adjust- ment of the interocular distance on rangefind- ers and heightfinders is limited. These limits These lists of qualifications were recom- mended to the Navy in March 1945. No action was taken on adopting them. Other Research In the course of the research just reviewed, a great deal of information was collected on the reliability of various visual and nonvisual tests and on the relation between different test scores. These data will be of interest to persons participating in future testing programs or test research. The remaining sections of this chap- ter therefore discuss these data along with accounts of the quantitative results of the vali- dation experiments. 83 DESCRIPTION AND RELIABILITY OF TESTS USED Before presenting the details of the valida- tion and testing programs, the various tests used in the course of the studies by the Height- finder Project and by Project N-114 will be described. The visual tests included the Shuron pupil- lometer, the NDRC interpupillometer, the modi- fied Massachusetts vision test kit, the Ortho- Rater industrial vision test, the projection eikonometer, the stereoscopic trainer M2, the vectograph-pursuit test, and the Dearborn- RESTRICTED 74 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTFINDER OPERATORS are 58 mm and 72 mm. Selection standards use the range from 60 to 70 mm to allow for errors of measurement and errors of rangefinder in- terpupillary scale calibration. Actually, not many men are eliminated on the basis of inter- pupillary measures. More Negroes than whites presentation is controlled by a curtain-slot arrangement which allows the subject to see only one row of letters at a time. The subject must read 80 per cent of the 25 letters correctly with each eye to qualify as having 20/20 vision. The reliability of this test at this 20/20 level is high, as indicated in the results of one ex- periment where 97 per cent of those who passed the test on its first administration also passed it on retest.28 Phoria measurements are taken by a Maddox rod technique. The subject views the phoria chart with his left eye. The chart has vertical and horizontal scales marked off in prism diopters. With his right eye, the subject sees a red streak, formed by the Maddox rod over his right eye and arising from a small electric light at the center of the chart. The position of the streak on the scales is the measure of phoria. The reliability of the test is comparable to that of other tests.28 For N = 152, the test- retest correlation for vertical phoria was r = -f-.59 (or = .05), and for lateral phoria was r = -j-.83 (ar = .03). 8.3.5 The Ortho-Rater Industrial Vision Tests The Ortho-Rater is a visual testing instru- ment developed by the Bausch and Lomb Opti- cal Company for use in industrial classification (see Figure 2). All tests on the instrument are given by viewing stereoscope-type slides. Among other measures of visual function which can be obtained with the Ortho-Rater are far vision tests of visual acuity (right eye, left eye, and binocular), phoria (vertical and lateral) and stereoscopic acuity. These tests were examined by Project N-114 for possible use in the selection of Navy rangefinder op- erators. The reliability of the Ortho-Rater tests as administered to Naval personnel by Naval ex- aminers was the first step in the investiga- tion.18 A test-retest study was carried out on a population of 234 recruits at Sampson. Results are given in Table 1. Retests were always given by a different examiner from the one who gave the original test. Figure 1. Distributions of interpupillary dis- tances for white and Negro soldiers. are eliminated on this basis, however.9 Distri- butions of interpupillary distance measures for white and Negro populations are shown in Figure 1. 8.3.4 Modified Massachusetts Vision Test Kit The equipment in this kit consists of a visual acuity chart and phoria test materials including a chart and Maddox rod spectacles. The kit was developed and introduced by Division 7.2-4 The visual acuity chart uses Snellen E’s. After the subject reads some larger letters, he is given 25 letters of the 20/20 size to read with his right eye and 25 to read with his left eye. Letter RESTRICTED DESCRIPTION AND RELIABILITY OF TESTS USED 75 HEADREST LOOP HANDLE CONTROLS VIEWING BOX POSITION FOR NEAR TESTS (AS SHOWN) OR FOR FAR TESTS LENSES LARGE DRUM WITH SLIDES FOR FAR VISION TESTS VIEWING BOX SMALL DRUM WITH SLIDES FOR NEAR VISION TESTS CRANK FOR TURNING DRUM TO CHANGE FAR VISION TESTS CRANK FOR TURNING DRUM TO CHANGE NEAR VISION TESTS Figure 2. Bausch and Lomb Ortho-Rater. View from examiner’s position. (Courtesy of Bausch and Lomb.) Test-retest correlations were lower than might otherwise have been expected (compare columns 1 and 3 in Table 1) because the men in the sample had already been selected to an unknown extent by the Navy on the basis of several visual functions. Therefore, the basic evaluation of the reliability of the Ortho-Rater tests had to be made in terms of the data sum- RESTRICTED 76 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTFINDER OPERATORS marized in the second column of the table— data showing the percentage of test-retest differences which were one score unit or less. As indicated, the Ortho-Rater measures are sufficiently precise and dependable. The only differences between test means and retest means which were statistically signifi- cant were those for right eye and left eye acuity. In both cases, the shift in measures was in the direction of poorer scores on the retest. The standard errors of estimate were small The degree of correspondence between the two tests is shown further by these correlation values. Visual acuity, right eye r = +.80, o,. = .02 Visual acuity, left eye r = +.81, or = .02 Lateral phoria r = +.54, or = .04 Vertical phoria r = +.44, or = .05 It may be noted that the difference between these coefficients of intercorrelation and the previously cited coefficients of reliability for Table 1. Reliability of the Ortho-Rater industrial vision tests. Results on 234 boots tested by a Laboratory test of 188 industrial employees by number of Navy examiners one examiner Test-retest % of men with test-retest difference of Test-retest correlation 1 or less correlation Visual acuity, far vision, both eyes +.49, ±.05 74 Visual acuity, far vision, right eye + .59, ±.04 75 + .89, ±.01 Visual acuity, far vision, left eye +.59, ±.04 62 + .83, ±.02 Lateral phoria, far vision +.83, ±.02 84 + .82, ±.02 Vertical phoria, far vision +.55, ±.05 96 + .62, ±.05 Stereopsis +.64, ±.04 55 enough, however, to ensure satisfactory pre- diction in spite of significant test-retest differ- ences between mean scores. The test of steroscopic acuity given on the Ortho-Rater is a static test. The subject views a series of numbers and indicates which one in a row of five appears nearer to him. There are only six observations of this sort for the sub- ject to make. Reliability of the test, as shown in Table 1, is .64, 8'3'6 Relation between the Massachusetts Vision Test and the Ortho-Rater Tests of Visual Acuity and Phoria The Massachusetts vision tests and the Ortho- Rater tests of visual acuity and phoria are sat- isfactory substitutes for each other. They have equivalent reliabilities, and scores on the two tests are closely related as shown in Table 2, which summarizes the results of a special com- parison study carried out by experienced test personnel at Fort Eustis, Virginia.28 the two sets of tests is an indication of the extent to which the two sets of tests measure different things. The difference is small except Table 2. The number and the proportion of subjects passing the Massachusetts and Ortho-Rater tests on visual acuity, lateral phoria, and vertical phoria for distant vision.* Distant vision Visual Lateral Vertical Item acuity phoria phoria N 576 288 288 Pass Mass. Test N 410 273 279 % 71.3 95.0 97.0 Pass Ortho-Rater N 419 264 268 % 72.8 91.8 93.0 Pass both N 393 259 266 % 68.3 90.0 92.4 Agreement, pass or fail N 533 269 273 % 92.5 92.5 94.6 ♦Passing scores, Mass.: 80% correct ±6 diopters diopter Ortho-Rater: 8 or more 2 to 13 4 to 7 in the case of lateral phoria where the differ- ence between the Massachusetts and the Ortho- Rater measurement methods may be of some importance. RESTRICTED DESCRIPTION AND RELIABILITY OF TESTS USED 77 837 The Projection Eikonometer Among the instruments invented by the Dartmouth Eye Institute to measure stereo- scopic vision is one called the projection eiko- nometer. This instrument, and in particular one test given with it, the stereo-vertical test, was further developed and applied by the Princeton University (Division 7) Field Lab- oratory at Fort Monroe, Virginia. Its eval- (see Figure 3). The two images are polarized so that when they are viewed through proper Polaroid filters, only one line is visible to each eye. The subject fuses these separately seen images and perceives a single line. During the projection eikonometer test, the two lines are rotated in opposite directions. This causes the perceived line to appear to tilt away from the vertical, either toward the subject or away from the subject. A motor drive returns the Figure 3. Arrangement of the projection eikonometer for testing a single subject. uation was continued by the Heightfinder Project N-114 under the Applied Psychology Panel. In the stereo-vertical test, two lines are projected simultaneously on a vertical screen lines toward the vertical. The subject presses a button at the instant the stereoscopically fused line appears to be vertical. The angle which the two images on the screen make with each other at the time of the subject’s response RESTRICTED 78 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTFINDER OPERATORS is recorded. If the subject responds on every trial with zero error (and if he has no cyclo- phoria), he will always respond when the two are made as the stereoscopically perceived line alternates in direction of movement. In each “run,” then, there are five times when the line Figure 4. Arrangement of the multiple projection eikonometer images on the screen are parallel and vertical. The complete test consists of three “runs” of ten observations each. Successive observations is tipped away from the subject initially and moves toward him during the trial, and five when it is tipped toward him initially and RESTRICTED DESCRIPTION AND RELIABILITY OF TESTS USED 79 straightens up by moving back. The subject is scored on the basis of the variability of his responses. Raw scores are converted to stand- ard scores by use of a conversion table. Details of the eikonometer testing and scor- ing procedures are given in references 6, 8, 19, and 20. A disadvantage of the original projection eikonometer test was that from 25 to 30 min- utes were required to test each subject, and only one subject could be tested at a time. The needed development of the projection eiko- nometer into a group testing instrument was begun under Division 7 and successfully con- cluded by Project N-114. The new instrument, known as the multiple projection eikonometer, permits six subjects to view the screen simul- taneously. The subjects sit in a small gallery with two rows of three seats each (see Figure 4). While all subjects do not see the screen from exactly the same angle and distance, it was demonstrated that test scores were not affected by these small differences in angle and distance.11 In the final form of the multiple in- strument, each subject has a noiseless key which he presses when the line appears to him to be vertical. The time of the true vertical and the instant at which each key is pressed are recorded on a tape recorder (see Figures 5, 6, and 7 for details of projector and recorder). With the multiple projection eikonometer, six subjects can be tested in from 40 to 50 minutes. A slightly longer time is spent on the instruction of subjects with the multiple in- strument than is required with the single instrument. Otherwise, the administration of the tests is the same. During the months of January and February 1943, while the multiple projection eikonometer was in operation at Fort Eustis, Virginia, arrangements were made to test a number of men on both the multiple and single instru- ments.11 As time in the Testing Center allowed, men were given both tests. One half of the men took the multiple test first (the MS group), and the other half took the single test first (SM group). The experimental data are summar- ized in Table 3. Means and standard deviations of the distributions of standard scores on the tests are given for the two halves of the popu- lation, SM and MS groups, and for both instru- ments. The standard errors of the means and the critical ratios of the differences between means are also given. All of the critical ratios were small. Scores on the single instrument averaged slightly higher than scores on the multiple instrument, but this difference was not statistically reliable. Table 3. Comparison of scores obtained on the single and multiple projection eikonometers. Instrument N= 99 SM group M=100 MS group Single eikonometer Mean = 112.01 cr = 17.8 C Mean = 113.97 :.R. =0.81 24> 25> 26< 28 on the intercorrelation between tests of stereoscopic vision. For purposes of direct comparison, the results of these studies are presented in Table 11. Intercorrelations of stereoscopic tests: dataf from Tufts College studies, Division 7.24- 28> 26 Tufts Vectograph- Botts Keystone Wulfeck trainer pursuit test test test M2 trainer (+.71) .47* .50* .22* .35 Tufts trainer ( + .55) .54* .34* -.11 Botts test ( + .79) .32 + .69* -.39 *Significantly greater than zero. fPopulation of 56 men (all 20/20 vision) and 32 women (no glasses). (Population smaller on Wulfeck test and vectograph test.) Reliabilities are given in parentheses. Tufts trainer, M2, and vectograph-pursuit were dynamic tests; Botts, Keystone, and Wulfeck tests were static tests. When a visual acuity standard of 20/20 in each eye (unaided) was applied, 27.6 per cent of the men secured a standard score of 110 or higher on the projection eikonometer. If the visual standard were dropped to include all those who make no more than one error on the Tables 4, 11, 12, and 13. The conditions of each of the experiments are summarized in the table footnotes. For a more complete description of those tests which are new to the reader, reference to the original reports is recommended. The tests RESTRICTED FUTURE WORK ON VISUAL TESTS 91 involved are of two types—“dynamic” tests, where the subject sees a target move or moves it himself, and “static” tests where the subject merely makes a judgment about the relative stereoscopic position of two or more figures. 89 FUTURE WORK ON VISUAL TESTS Neither the medical interest nor the scientific interest in problems of vision has yet led to the production of a completely satisfactory set of Table 12. Intercorrelations of stereoscopic testsf: Fort Eustis data for 192 men with vision of 20/20 or better.28 M2 VP DJ W O-R Keystone Projection eikonometer .42* .23* .38* - .29* - .39* -.34* M2 trainer (M2) .36* .41* - .49* - .38* -.41* Vectograph-pursuit (VP) .31* - .14* - .31* - .30* Dearborn-Johnston (DJ) -.31* - .29* -.28* Wulfeck (W) .29* .33* Ortho-Rater (O-R) .47* ♦Significantly greater than zero. fProjection eikonometer, M2 trainer, vectograph-pursuit, and Dearborn-Johnston tests were all dynamic. Wulfeck, Ortho-Rater, and Keystone tests were all static. Negative coefficients in the dynamic vs static block of Tables 12 and 13 show positive relations because the dynamic tests were graded as “error” scores and the static were graded as “right” scores. The correlations which are significantly greater than zero (confidence level of 5 per cent) are marked with an asterisk in the tables. The important fact about the intercorrela- tions shown in Tables 4, 11, 12, and 13 is that, although many of them are significantly visual tests. Some visual functions can be meas- ured fairly reliably, but for other functions (see also Chapter 9) the best available instru- ments provide scores of very low reliability. Tests which purport to measure the same func- tions show low correlations with each other. Table 13. Intercorrelations of stereoscopic testsf: Fort Eustis data for 96 men with less than 20/20 vision.28 M2 VP D-J W O-R Keystone Projection eikonometer .37* .46* .34* - .09 - ,11 - .14 M2 trainer (M2) .53* .21* - .21* - .27* -.42* Vectograph-pursuit (VP) .36* - .19 -.21* - .31* Dearborn-Johnston (DJ) - .51* -.06 -.12 Wulfeck (W) .10 .39* Ortho-Rater (O-R) .29* *Significantly greater than zero. fProjection eikonometer, M2 trainer, vectograph-pursuit, and Dearborn-Johnston tests were all dynamic. Wulfeck, Ortho-Rater, and Keystone were all static. Negative coefficients in the dynamic vs static block of Tables 12 and 13 show positive relations because the dynamic tests were graded as “error” scores and the static were graded as “right” scores. greater than zero, none is high enough to per- mit reliable prediction from one test to another. From this and the argument already presented in Section 8.3.11, it is safe to conclude that the data do not indicate that any one of the tests is a substitute for another. Thus, it is apparent that validation results obtained for one test in a selection experi- ment cannot be used to indicate what might happen in a similar experiment with another test. A fairly extensive program of research on visual requirements and visual tests would have a very good chance of producing results of great value to the Army and the Navy. The research should include: 1. A thorough experimental and statistical analysis of the various visual functions and abilities and of the relations between them. 2. A detailed study of the requirements of a number of representative military tasks in terms of each type of visual ability. RESTRICTED 92 SELECTING STEREOSCOPIC RANGEFINDER AND HEIGHTFINDER OPERATORS 3. The construction of new tests and the im- provement of existing tests to measure each type of visual ability. Several tests for each type should be available. By the regular pro- cedures of test standardization and validation, the best tests could then be selected and vali- dated. Because many of the visual tests in use today were devised from a medical point of view and were intended to measure the degree of visual impairment or defect, they have not always proved too satisfactory as predictors of suc- cess in jobs which involve visual ability. The tests and testing procedures described in this chapter were developed especially for use in selecting good men. Approaching the whole visual testing problem from the standpoint of predicting visual proficiency may lead gener- ally to tests of greater usefulness in military selection. The effort is well worth making. RESTRICTED Chapter 9 SELECTING NIGHT LOOKOUTS William E. Kappauf, Jr.a Summary The Applied Psychology Panel and the National Research Council’s Committee on Human Aspects of Observational Procedures cooperated with other groups in the develop- ment of two adaptometers, the NDRC adaptom- eter, Model III, and the modified Rostenberg adaptometer. Two studies on the prediction of night look- out performance both showed that the tests of night vision being used had low reliability and very low validity. 91 THE PROBLEM OF SELECTING LOOKOUTS For 99.9 per cent of the time the job of night lookout is a thankless one. It is tedious, unin- spiring, poorly rewarded work. Men assigned to the job are often those left over when all other billets have been filled. They cannot be rated as long as they stay at the job. These conditions have meant that, during the war, night lookouts were generally chosen by negative selection rather than by positive selec- tion. Only one particular capacity has been sought in the lookout—good night vision, or at least evidence of no night blindness. This in itself might be an adequate selection criterion were it not for the fact that most night vision testing devices, usually called adaptometers, are of very low reliability. Men taking the test usually improve their scores when retested and do not rank in the same order of ability on the second test. It is therefore not surprising that a great amount of research time during the war should have been devoted to the development of a more satisfactory and dependable adaptometer and to answering the fundamental question as to whether an adaptometer is a good selection device for night lookouts. The Applied Psy- chology Panel participated in this work. It assisted in the design of two adaptometers and conducted field experiments in which tests of night vision were evaluated for their useful- ness in predicting night lookout performance. 92 THE NDRC ADAPTOMETER The National Research Council’s Committee on Human Aspects of Observational Proce- dures in cooperation with other OSRD sections and the Navy developed the NDRC adaptom- eter, Model III.1 This adaptometer is similar to the radium plaque adaptometer officially adopted for use by the U. S. Navy, except in its source of illumination. An incandescent bulb rather than a radium plaque provides the neces- sary illumination. The man being tested fixates a small red cross and reports the orientation of a peripherally located T-shaped figure which is flashed on for brief intervals. The observa- tion distance is five feet. Testing can be carried out at any of five different brightness levels (see Figures 1 and 2). This instrument was not adopted for routine use by any Service. 9 3 THE MODIFIED ROSTENBERG ADAPTOMETER In response to a request from the Techni- cal Section, Classification and Replacement Branch, Office of the Adjutant General, U. S. Army, a representative of the Applied Psy- a This chapter is based on the work of NDRC Project N-115 and work done by the National Research Coun- cil’s Committee on Service Personnel—Selection and Training, The Applied Psychology Panel was only one of sev- eral groups seeking better methods of selecting men for duty at night. Others working on this problem included the Adjutant General’s Office, the Army Air Forces, the Armored Medical Research Laboratory, and the Medical Research Laboratory, U. S. Submarine Base, New London, Connecticut. Some of the reports from these other agencies claimed more positive results than those described in this chapter. RESTRICTED 93 94 SELECTING NIGHT LOOKOUTS chology Panel Project N-115 went to Camp Blanding, Florida, to assist in testing an adaptometer designed by Major Rostenberg. Only preliminary tests were run, but as a result Medical Research Laboratory, Bethesda, Mary- land, for test and further study. No report on the results of this test has been received. The modified Rostenberg adaptometer has a Figure 1. NDRC adaptometer, Model III, front view. of this experience in night vision testing, the personnel of Project N-115 later constructed two modified and improved models of the Rostenberg adaptometer.5 One of these modi- fied instruments was transferred to the Naval radium plaque as its source of illumination. Test objects of any shape or size up to 16 inches in diameter can be shown in any of 8 positions. The adaptometer is used at a distance of 20 feet. Thus, the subject uses distance vision as RESTRICTED THE MODIFIED ROSTENBERG ADAPTOMETER 95 Figure 2. NDRC adaptometer, Model III, rear view. RESTRICTED 96 SELECTING NIGHT LOOKOUTS demanded in the lookout situation. Eight men can be tested at one time, which makes it pos- sible to increase the number of trials given each subject without increasing the total testing time. The adaptometer is simple in construction and operation and has easily controlled illumi- nation. It has but two disadvantages: it requires a fairly large, dark room for testing and it is larger than most other available adaptometers. 9 4 EXPERIMENTS ON THE PREDICTION OF NIGHT LOOKOUT PERFORMANCE Study of the validity of adaptometers and night vision tests was undertaken by the Applied Psychology Panel at the request of the Navy. The investigation was carried out by Project N-115 working in cooperation with the Medical Research Laboratory at the Submarine Base, New London, Connecticut.2 Data were sought on the following points: (1) the relia- bility of a group of night vision tests, (2) the consistency of night lookout performance, or the reliability of measures of such perform- ance, and (3) the relations between night vision test scores and night lookout performance. 9-41 Reliability of Night Vision Tests The tests which were used in the experiment were the following: 1. The radium plaque adaptometer. This in- strument was in use as the official U. S. Navy adaptometer. From a distance of five feet the subject reported the orientation of a T figure on a 5-degree white field of low illumination. The score was the number of correct observa- tions. 2. The NDRC adaptometer, Model III. This test, as described in Section 9.2, was the same as the radium plaque adaptometer test so far as the man being tested was concerned. A slightly lower stimulus brightness level was used with this instrument than the level pro- vided in the radium plaque adaptometer test. 3. The clockface adaptometer. This was a modified form of the method of testing night vision used by the amphibious forces. The sub- ject was scored in terms of the distance from which he could correctly report the orientation of a white hand on a clockface which had white markers at 12, 3, 6, and 9 o’clock. 4. The Tufts-SDS test of night vision. This was a visual acuity test (Snellen E’s) at low illumination. 5. A Purkinje test. This was a test of adap- tation time, the time required for the subject to adapt sufficiently to see one of four test areas of low illumination. 6. The Hecht-Schlaer adaptometer, RCN Model. This instrument was the official adap- tometer of the Royal Canadian Navy. It meas- ured the subject’s low brightness threshold for a 1/5 second flash. 7. A test of visual-motor coordination car- ried out at low illumination. This test was de- veloped by K. U. Smith for the Committee on Aero-Medical Research. 8. The night lookout training stage. Per- formance on the training stage was scored in terms of the voltmeter setting necessary to see a ship model on the stage. Records were taken at the end of each of three training sessions. 9-12. Four personality tests. 13. Score on the verbal portion of the Gen- eral Classification Test. These tests were administered to a group of 150 seamen, none of whom had had experience as a lookout. Shortly after their arrival at the submarine base the men were given the tests listed above. The testing required three days. They were then given a regular three-day course of instruction in night lookout training. During the following three days the entire group was retested on the series of tests. Data on reliability of the tests may be sum- marized by saying that none of the visual tests had a reliability coefficient high enough to be satisfactory for efficient selection, Test-retest correlations for four of the tests are shown in Table 1. Reliability values for the other visual tests, where the distributions of test scores were par- ticularly skewed, were obtained by the chi- square method. Four-cell tables were set up with various cutting scores applied to divide passing and failing groups on the first and RESTRICTED EXPERIMENTS ON THE PREDICTION OF NIGHT LOOKOUT PERFORMANCE 97 second tests. Values of chi-square significant at the 1 per cent level of confidence or better were obtained at some cutting level for all the tests. It was frequently true, however, that this confidence level was obtained only for cutting scores which would be impractical for use in liable. Reasons for expecting unreliability in lookout performance scores are many. They in- clude trial-to-trial variation in target size, target speed, target angle, and sky brightness, the small number of trials, the relative inex- perience of the men, and the small difference in ability found among the 58 men. 9 4 3 Relations between Night Vision Tests and Lookout Performance Ten of the tests were useless in predicting the night lookout performance of the 58 men in the criterion group. These ten tests were the radium plaque adaptometer, the Purkinje test, the Hecht-Schlaer adaptometer, the Smith vis- ual-motor coordination test, the night lookout training stage, the four personality tests and the verbal score on the General Classification Test. The three remaining tests were positively related (according to chi-square measures) to night lookout performance. These tests were the NDRC adaptometer, the clockface adap- tometer, and the Tufts-SDS test. Of these the clockface adaptometer and the Tufts-SDS test were valid only for cutting scores set high enough to fail more than 20 per cent of the group. They were, therefore, deemed imprac- tical for use at the moment. The NDRC adap- tometer, which differentiated between good and poor lookouts at a level which did not fail so many men, was at best a poor predictor of night lookout performance. It should be ob- served that although scores on this instrument were positively related to lookout performance, scores on the radium plaque adaptometer were not. Since these two instruments are very sim- ilar in nature and show a relation to one an- other which is nearly the same as the reliability of each, it is believed that the difference in pre- dictive value between the two may have been due to the small number of men in the criterion group who failed the radium plaque adaptom- eter test and may, therefore, have been a chance difference. In summary of this experiment, then, it was concluded that none of the seven tests of night vision and none of the six other tests tried Table 1. Reliability of night vision tests. (N — 150) Test-retest Test correlation Clcckface adaptometer (from Amphibious Forces) + .27 Tufts-SDS test of night vision + .19 Purkinje test (adaptation time) -j- .53 Hecht-Schlaer adaptometer (Royal Ca- nadian Navy Model) + .42 Service selection; i.e., cutting scores which failed as many as 40 per cent of the total popu- lation. 942 Consistency of Night Lookout Performance In the study of night lookout performance, 40 of the subjects were taken aboard ship each night and records were taken of their ability to pick up and report an approaching target ves- sel. Several target runs were made during each watch. A watch lasted 30 minutes to an hour and each man stood three watches. Target range and bearing information were secured by radar. Because of interference from airplane search- lights the data for only three nights of testing could be used. Several other factors combined to reduce the size of the test population still further. There were finally only 58 lookouts on whom useful performance scores were avail- able. In order to obtain a measure of the reliability of lookout performance, comparisons were made between the average range at which a partic- ular man spotted a target on his first, third, fifth, etc., trials and the average range at which he spotted the target on his second, fourth, sixth, etc., trials. The product-moment correla- tion between the odd-trial average range and the even-trial average range for the 58 men was +.56. The performance was not very re- RESTRICTED 98 SELECTING NIGHT LOOKOUTS proved to be useful as methods of predicting the quality of shipboard night lookout perform- ance. Additional data on the efficiency of the radium plaque adaptometer in selecting night lookouts was obtained in another experiment4 carried out as part of a shipboard study of night lookout performance which is described in detail in Chapter 6 of Volume 2, Applied Psychology Panel. Unfortunately, the radium plaque adaptometer does not yield scores that can be correlated easily with lookout perform- ance scores. In the first place, it gives an all- or-none type of score: the man either passes or fails. In the second place, the man is tested at only one brightness level. Thus, between two men who pass or between two men who fail there may be very large differences in ability at the tested or at untested levels of illumina- tion. It is possible, nevertheless, to make some rough comparisons between radium plaque adaptometer scores and the performance scores obtained aboard ship. Of the 114 men who served as subjects in this experiment, 63 had been tested on the radium plaque adaptometer by one of the ship’s pharmacist’s mates, who was a qualified oper- ator of the adaptometer. Of the 63, 50 had passed and 13 had failed the test. When the lookout performance of these men was meas- ured in terms of ESF scores (equivalent square feet of target visible at 1,000 yards; see Chap- ter 6, Volume 2, Applied Psychology Panel for complete definition), it was found that the aver- age score for those who had passed the radium plaque adaptometer test was 64.7 ESF, a score men. Ten of the 13 men who failed the adaptometer test had better lookout scores than the mean score of those who passed the test. The data were considered in another way: the very best lookouts and the very poorest lookouts as measured by ESF scores were com- pared with respect to adaptometer perform- ance. The results of this comparison are sum- marized in Table 2. The difference in lookout scores between these two groups was clearly significant in a prac- tical sense as well as in a statistical sense. The 12 best men could, on the average, see a ship about four times as far away as the 12 poorest men. But the two groups were equivalent with respect to radium plaque adaptometer perform- ance. This comparison showed, as did the pre- vious experiment, that the relation between scores on the radium plaque adaptometer and the performance of men as night lookouts is essentially zero. Recognition of this fact ap- peared in the “Manual of Procedures for U. S. Naval Classification Centers,” NavPers 15082, dated February 1945, page 150, which contains these instructions: This test (Radium Plaque Adaptometer) is administered under the cognizance of the Bureau of Medicine and Surgery. The Bureau of Naval Personnel has directed that scores should not be used in deter- mining qualifications for any school or rate, including Armed Guard. Scores for this test should not be recorded on the Enlisted Personnel Qualifications Card. 9 5 CONCLUSIONS ON LOOKOUT SELECTION Failure to achieve satisfactory methods for selecting night lookouts hinges on at least three factors: the unreliability of night vision tests, the unreliability of lookout performance measures, and the probable importance for lookout performance of traits other than those measured by the tests used. Either a different kind of scotopic acuity is involved in the two sets of measurements (adaptometer and look- out situation), or the lookout scores are pri- marily measures of motivation and application Table 2. Comparison of radium plaque adaptom- eter scores for best and poorest lookouts. Group Mean look- out score Radium plaque adaptometer record Pass Fail No record 12 best men 12 ESF 8 1 3 12 poorest men 196 ESF 7 1 4 about equal to the average for all the 114 men. The average score for those who had failed the test was 43.9 ESF, a score which was better than the average score for the 114 RESTRICTED CONCLUSIONS ON LOOKOUT SELECTION 99 to the job rather than measures of visual abil- ity. Only future research can determine which is the case. Toward the development of more efficient lookout performance much can be done by in- creasing the prestige of the lookout, by the better training of lookouts and lookout officers, by formalizing the lookout organization aboard ship, and by making routine inspections of lookout procedures.3 These matters are dis- cussed in more detail in Chapter 6 of Volume 2, Applied Psychology Panel. RESTRICTED Chapter 10 SELECTING NAVY TELEPHONE TALKERS Dael Wolfle a The particular impetus for the project was a request for the immediate classification of men on the USS New Jersey. After work on this ship was finished, the speech interview as a means of rating telephone talkers was adopted at other Navy installations. 102 A SPEECH INTERVIEW FOR THE SELECTION OF TELEPHONE TALKERS A speech interview1 to rate the potential ability of men as shipboard telephone talkers was developed. During the interview the men pronounced numbers, repeated commands, and gave an extemporaneous description. All com- mon American speech sounds were elicited dur- ing the interview. Because of the urgency for immediate classi- fication of men for the particular ship, the USS New Jersey, on which speech work began, it was impossible to take the time necessary to validate the speech interview as a selection method. Instead, methods of eliciting and rating speech were adapted from college speech-pro- ficiency interviews. The interview was modified during preliminary exploratory work with 290 recruits selected at random from the Receiving Ship at the Philadelphia Navy Yard. Officers concerned with classification had stipulated that the speech rating should divide the enlisted personnel into three categories as talkers: well qualified, qualified, and not quali- fied. The speech interview was intended to meet this requirement. It was included as a part of the classification process and preceded the gen- eral classification interview. In this way a man’s speech rating was available to the interviewer when the man’s duty assignment was decided upon. In April 1943, project members began ad- ministering the speech interview to the pros- pective crew of the USS Neiv Jersey. Speech ratings were obtained for 2,114 enlisted men Summary A speech interview to rate the potential abil- ity of men as shipboard telephone talkers was developed. The interview was conducted over Navy sound-powered telephones. It in- cluded the pronunciation of numbers and of all common American speech sounds, repetition of commands, and extemporaneous description. Men interviewed were rated as “well qualified,” “qualified,” or “not qualified.” A training course on the administration of the speech interview and the rating of speech quality was developed and taught to a number of classification interviewers. The speech interview was adopted by the Bureau of Naval Personnel for routine use in the classification of enlisted personnel. 101 INTRODUCTION A great deal of the communication within a Navy ship is by means of sound-powered tele- phones. The term “sound-powered” is used be- cause the phones are activated by the energy of the speaker’s voice without an external cur- rent source. These phones have the obvious advantage of allowing communication to be independent of the ship’s power sources. But they cause difficulties in communication: fidel- ity is poor and the energy level is low. Recog- nizing these disadvantages, the Navy requested the Committee on Service Personnel in March 1943 to assist in the selection and training of those members of a ship’s company who could talk most effectively over the sound-powered phones. The request was accepted and assigned as Project N-109 to the Psychological Corpora- tion for solution. The work on selection of telephone talkers is described in this chapter; that on training of telephone talkers in Chapter 10 of Volume 2, Applied Psychology Panel. a This chapter is based upon the work of Project N-109. 100 RESTRICTED TRAINING OF SPEECH INTERVIEWERS 101 on a seven-point scale: superior, very good, good, average, poor, very poor, and inferior. These ratings were immediately evaluated by the judge into the three categories requested by the classification officers. The testing program, in addition to supplying the data desired by the ship’s officers, permitted the study of (1) the performance of skilled judges in making rating judgments, (2) the frequency of various types of speech deviations, and (3) the rela- tionship of the several aspects of speech to in- telligibility over voice communication systems. While the testing program was in progress, 419 men who had been given the speech inter- view were also tested for listening and memory- span abilities by representatives of the Harvard Psycho-Acoustic Laboratory. An investigation of the relation between listening ability, mem- ory span, and speech ability was thus made possible (STR of NDRC Division 17, Volume 8). The following conclusions appeared to be warranted.1 1. A speech rating can be assigned to each member of a ship’s enlisted personnel through a speech interview at the time that he is classi- fied for ship’s duty. 2. The original seven-point speech ratings fell into an essentially normal distribution. 3. Neither sea experience, telephone-talking experience, nor education is closely related to an enlisted man’s speaking ability as judged by this method. 4. Speech traits that accompany limited in- telligibility are: poor articulation, foreign or regional dialect, too weak a voice, poor control of intensity, and a voice quality heard as harsh, hoarse, or husky. 5. All factors considered, those speaking the general American dialect were judged to be slightly more intelligible than those speaking the other two major dialects, southern or eastern, 6. Expert judges making independent rat- ings agree closely in evaluating intelligibility. 7. Expert judges agree more than 90 per cent of the time in placing men in one of the three categories of well qualified, qualified, or disqualified. 8. Speaking ability is not highly related to memory span or listening ability. 10 3 TRAINING OF SPEECH INTERVIEWERS3’ 4> 5> c’7 Results obtained in the selection of telephone talkers for a number of individual ships led to a demand for wider use of the speech interview. Expert speech men were not numerous enough to interview all recruits. It therefore became necessary, first, to determine whether or not the speech interview could be given successfully by classification interviewers2 and, second, to develop a course for training classification in- terviewers to administer the speech interview. Project members prepared a short manual on the selection of telephone talkers.4 This man- ual, and a college textbook on voice and artic- ulation (Fairbanks: Voice and Articulation Drillbook), were used as texts in a course given by project personnel to train classifica- tion interviewers to administer the speech inter- view.5 In addition to their reading, future inter- viewers also engaged in a number of types of drill. They were encouraged to practice speak- ing exercises during their free time. Each stu- dent was provided with five-minute recordings of his speech at the outset of the course and at the end of the course. Two Mirrophones were available all the time for practice. Class periods included lectures on speech; Mirrophone drill; practice in judging recruits’ speech in face-to- face situations, from recordings, over tele- phones, and over ship’s dial telephones; and practice in making repeated judgments of merit of recorded speech interviews. Results obtained with this course indicated that classification specialists could be trained to administer the speech interview satisfac- torily. The student subjects improved in their ability to evaluate speech, and as the training progressed their judgment coincided more and more closely with that of expert judges and with the average of group judgments. The re- sults were regarded as tentative, as it was felt that more effective training methods might be developed in the future. Typical correlations of the ratings made by petty officers trained to administer the speech interview with ratings of the same men made by an experienced speech teacher were about RESTRICTED 102 SELECTING NAVY TELEPHONE TALKERS .80, This level of agreement was considered sufficiently high to justify using classification petty officers for routine administration of the speech interview. 104 LIMITATIONS ON THE USE OF THE SPEECH INTERVIEW The ultimate fitness of a telephone talker depends, not only upon the intelligibility of his speech, but also upon a fairly specialized knowl- edge of equipment, organization, and proce- dures, as well as upon his attitudes, sense of responsibility, and behavior under stress. Thus a rating of “qualified” or “well qualified” on the classification card should not be interpreted to mean that a man has been certified as fully capable of performing all the duties of a tele- phone talker. It means merely that the man’s speech intelligibility is passable; it is such that with proper training and experience he can become a qualified or well qualified talker, 10.S NAVY ADOPTION OF THE SPEECH INTERVIEW As a result of the development of the speech interview and the studies in the training of classification specialists to administer it, the Navy began to use the method at Naval Oper- ating Base, Norfolk, Virginia. Over 20,000 men were rated as telephone talkers at that station, and the ratings were noted on their qualifica- tions cards. The speech interview was later ap- proved as a standard classification procedure by the Bureau of Naval Personnel (letter, Pers- 6372-MS-4, RS-F-4d, dated April 15, 1944). It became a routine part of the Navy classification and selection practice in a number of stations on both the east and west coasts.8 RESTRICTED Chapter 11 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION John L. Kennedy11 Summary COOPERATIVE WORK by the Navy and NDRC developed mass classification procedures for the crew of the USS New Jersey, for the Am- phibious Training Command of the U. S. Atlan- tic Fleet, and for the crews of destroyers being manned on the west coast. The principles and procedures developed were adopted for use throughout the Fleet. The objectives of classification are to assign each man to the job he can do best and to divide available manpower equitably among the vari- ous specialties and among the various units of a military force. Requirements for the success of a classifica- tion program are described. A detailed knowl- edge of the requirements of each military spe- cialty is necessary. Tests used to determine a man’s qualifications should be as independent of each other as possible; they must be vali- dated. Classification data must be kept up to date and available; procedures must be stand- ardized ; and the programs should be under the supervision of classification specialists. Several devices to aid in securing reliable information for use in classification were de- veloped and installed on several ships. A me- chanical device which does much of the work of combining and weighting classification data is described. 111 INTRODUCTION World War II demonstrated that the concept of an untrained citizen army, springing to arms overnight in a national emergency, is no longer tenable. Wars are fought more and more by specialists. Even the infantryman, the tradi- tional “cannon fodder,” emerged from the war as a specialist in the use of a complicated array of weapons requiring a considerable period of training to master. The large number of new weapons put further stress on specialization in the military forces and further complicated the training program. The growing specialization and the consequent time required for training in World War II emphasized the importance of scientific classification, allocation, and as- signment of manpower. Discussion of these topics is divided between this and the two following chapters. This chap- ter gives a brief account of the development of mass classification procedures by the Navy and NDRC and then presents some general princi- ples of good classification procedure and some devices to aid classification. Chapters 12 and 13 present in greater detail improvements which could be made in recruit and advanced classi- fication programs. 112 NAYY-NDRC DEVELOPMENT OF MASS CLASSIFICATION PROCEDURES Early in 1943 the Applied Psychology Panel was requested to assist in the classification and assignment of the crew of the USS New Jersey. Growing out of this cooperative Navy-NDRC effort came two requests, (1) to develop classi- fication procedures for the Amphibious Train- ing Command, U. S. Atlantic Fleet, and (2) to aid in the development of classification proce- dures for men assigned to duty on new destroy- ers that were being commissioned on the west coast. Simultaneously the Bureau of Naval Person- nel was extending and improving its work in the classification field. The success of these programs led to the adoption on a fleet-wide basis of the principles and procedures developed and tested in earlier programs. a This chapter is based upon the experience of the Applied Psychology Panel and the work of several of its projects which cooperated with the Bureau of Naval Personnel in establishing classification procedures. RESTRICTED 103 104 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION 11-21 The USS New Jersey Approximately 2,600 men made up the crew of the USS New Jersey, the first of this coun- try’s 45,000-ton battleships. Two-thirds of the men had never been to sea before their assign- ment to the New Jersey. The need to get the ship ready for fleet duty as quickly as possible was great. To expedite that process, the Bureau of Naval Personnel and NDRC were requested to analyze the nature of the ship’s many duties, to test and interview each man, and to recom- mend an assignment for each which he could learn most rapidly and in which he could work most efficiently. In addition to recommending assignment for each member of the ship’s company, a mechan- ical method for the rapid identification of per- sonnel with special skills and aptitudes was installed on the Neiv Jersey. Classification of the crew of the New Jersey was an emergency effort, done with little time for detailed preparation and without oppor- tunity for validating all the tests used. But it was a successful effort. Fewer transfers to new duty occurred on the New Jersey than were necessary on other ships. Men learned their duties more rapidly. Fewer disciplinary prob- lems arose. The Commanding Officer summar- ized these advantages in reporting: “The initial placement of personnel has definitely contrib- uted to the apparently extraordinary rapid progress made during the shakedown period by the crew of this vessel.” (Letter BB62/P16-3, Serial 570, August 31, 1943, to Chief of Naval Personnel). 11-2-2 Atlantic Fleet Amphibious Training Command The second request for assistance in devel- oping classification procedures for large groups of men came in a request from the Training Officer of the Amphibious Training Command, U. S. Atlantic Fleet, to assist in establishing a model classification program and to carry out research to improve the classification proce- dures already set up. Project N-117 with the State College of Pennsylvania was activated by the Applied Psychology Panel in order to meet this request. The project, in cooperation with officers from the Bureau of Naval Per- sonnel and officers from the Amphibious Train- ing- Command, was given the opportunity to work out broad, general procedures for the allocation of manpower as well as the methods for classification into specific job assignments. The Amphibious Training Program was a new development made necessary by the condi- tions under which World War II was fought. No blueprints existed for amphibious organ- izations; even the type of craft used was a matter of continuous development. The variety of specialized jobs to be done on board these craft meant that considerable time could be saved by the application of scientific procedures for handling the thousands of inexperienced officers and men who passed through the train- ing program. The intensive work of the project and its collaborators resulted in the adoption of a “Manual of Classification Procedures for Amphibious Training Bases”13 at all the am- phibious training bases on the east coast. Emphasis should be given the importance of the work in the Amphibious Training Command as it bears on the overall problem of mass allo- cation, classification, and assignment of per- sonnel. A thoroughgoing attempt was made in this work to control the input to the training program and in the assignment process to “balance” crews in terms of abilities and apti- tudes. In other words, instead of leaving the “raw material” for the classification and as- signment process to chance, planned use was made of test scores, previous job history, spe- cial skills and potentialities (determined by testing and interviewing before entrance into the training program) so that the output for assignment to specific ships could be controlled. In the assignment to ships after training at the amphibious training bases it was discovered that an enterprising officer was able to round up a crew which consisted entirely of “top” men. This meant that the next crew had to be made up from a pool with the top removed by the previous assignment. Such a process, when continued, resulted in marked unevenness in the ability of various crews. From the standpoint of the officer in charge of a particular ship, it was obviously desirable RESTRICTED GENERAL PRINCIPLES OF GOOD CLASSIFICATION 105 to have as good a crew as possible. But there were not enough good men to provide first-class crews for every ship. From the standpoint of the Service as a whole it was desirable to divide the good men among the ships, to “balance” the crews. One or two incompetent crews could have a disastrous effect on the efficiency of an entire attacking force if they failed to work properly in an attack landing. The principle of “balancing” and apportion- ing available brains and ability fairly among the crews had its initial full-scale trial in the work of this group. 11-2-3 Classification of Destroyer Crews Project N-116a with Stanford University was activated as the result of a request from the Commander, Fleet Operational Training Com- mand, Pacific, to assist in developing adequate techniques for classification for gunnery billets on all new destroyers manned on the west coast.11 The selectometer, a mechanical aid in classification and assignment of large groups of men, was one contribution of this project. It will be described in Section 11.4.4. Its devel- opment and use illustrate a second principle of mass classification, namely, the utilization of mechanical sorting, weighting, and summing procedures to make possible the rapid and standardized utilization of information about personal characteristics in classification and assignment. n.2.4 Fleet-wide Adoption of Mass Classification Procedures The Commander-in-Chief, United States Fleet, basing his directive on the success of the programs described above, on February 16, 1945 directed the Chief of the Bureau of Naval Personnel to develop a similar classification program for fleet-wide application (second en- dorsement to ComPhibTraLant Itr. FE 25/A7-3, Serial 0043, dated January 24, 1945). 113 GENERAL PRINCIPLES OF GOOD CLASSIFICATION The success of a military force depends upon the ability and training of the units of which it is composed. The force as a whole will nor- mally be most successful if the units are bal- anced in terms of ability and training. There are, therefore, two primary objectives of a classification program. One is to assign each man to the duty in which he can be most useful. The other is to distribute the available supply of ability evenly among the units composing a force. The classification experience of World War II indicates that realization of these two objectives requires observance of the following general principles. 1. Detailed knowledge of requirements for each military specialty is necessary. 2. Aptitude measures should be independent. 3. Aptitude tests must be validated. 4. Classification data should be constantly available. 5. Classification procedures should be stand- ardized. 6. Classification should be supervised by clas- sification specialists. These six general principles are discussed in the following sections. 11-31 Detailed Knowledge of Require- ments for Each Military Specialty Is Necessary Determining the requirements of a military specialty requires an analysis of the job itself by the techniques described in Applied Psychol- ogy Panel, Volume II, Chapter 14, in order to secure a detailed account of what is done. This analysis makes it possible to determine the mental and physical qualifications necessary to perform the duties of the job. It is sometimes fairly easy to determine the necessary qualifications. For example, the fact that the distance separating the eyepieces of a heightfinder can be adjusted through a fixed and limited range determines one necessary characteristic of a heightfinder operator: his interpupillary distance must fall within the range of possible instrument settings. More frequently it is necessary to determine empirically the exact requirements of a job. Such an empirical study is desirable both be- cause it provides information on essential RESTRICTED 106 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION qualifications and because it avoids the waste involved in setting too high selection require- ments. If selection standards are set too high, time is wasted in locating men who qualify, and the men selected are made unavailable for assignment to duties in which their high quali- fications would be more valuable. It is not difficult to select a group of men who will be successful in the Signal Corps, in Naval Aviation, or in some other specialty. It is more difficult, and in the long run more important, to determine the needs of each specialty, to assign each man to the specialty in which he will do best, and thus to divide the able men equitably among the various specialties. Aptitude Measures Should Be Independent If aptitude tests which purportedly measure different aspects of ability show high correla- tions with one another, they are measuring some general type of ability and failing to do the job for which they were intended; they will select men of generally high ability rather than picking out those high in the particular abilities needed in a given job. For example, if mechan- ical ability is important in one job and verbal ability in another, men cannot be selected effi- ciently for the two jobs unless the mechanical and verbal tests are relatively independent of each other. Aptitude measures should be as independent of each other as possible. 11-3-3 Aptitude Tests Must Be Validated No aptitude test can be used with complete confidence until it has been validated in accord- ance with the procedures described in Chapters 14 and 15. In wartime, tests must sometimes be used before validation is complete. It is then necessary to rely on the judgment of a test expert in deciding whether or not to use a test for classification purposes. The test expert may be wrong, as may an engineer; some tests have been worthless, and some bridges have col- lapsed. But when validation data are not avail- able, the test expert’s judgment is the best available basis for deciding what tests to use. The U. S. Navy Basic Classification Test Bat- tery, for example, was put into use before being validated. Later evidence showed it a valid battery and confirmed the judgment of the offi- cers who decided upon its immediate adoption. 113-4 Classification Data Should Be Constantly Available Classification data must be available in order to be used. The statement is so self-evidently true that its inclusion here should not be neces- sary. But there are so many ways in which classification data, carefully secured at one time, may become unavailable when needed that the principle demands emphasis.6 Classification data may become unavailable through any of these common means: 1. Failure to enter test scores and similar information on the man’s record card (WD AGO Form 20 in the Army; Q-card in the Navy; or their successors). 2. Failure to provide for recording school grades or other measures of demonstrated pro- ficiency. 3. Failure to transfer record cards when a man is reassigned. 4. Failure to provide proper mechanical de- vices for storing, sorting, and using record cards. 5. Failure on the part of an officer in charge to understand the nature and use of classifica- tion data supplied to him so that records are not properly preserved and hence are not there when wanted. These failures can all be avoided by proper indoctrination of officers and by inspection checks to determine that personnel policies are complied with. 11-3-5 Classification Procedures Should Be Standardized Many classification interviewers and person- nel assignment officers are required to keep the classification machinery operating. Their work must be done in accordance with established RESTRICTED DEVICES TO AID MILITARY CLASSIFICATION 107 and standardized forms, procedures, and rules. If all officers and men responsible for classifi- cation were highly competent experts in their tasks, rules could be fewer; but since they are not and in a large military organization never will be, standard procedures are necessary. Standardized procedures may result in some misassignment of men, but for the great bulk of the men being classified the total amount of misassignment will be smaller than it would be if each interviewer had complete freedom of choice in making his recommendations. The selectometer (Section 11.4.4) is one example of mechanical standardization of the task of recommending assignments. The superiority of test scores to interviewers’ judgments in selecting men for school training (Section 1.5.3) illustrates another aspect of the advan- tages of standardized classification regulations. There will of course always be some men whose qualifications are so unusual, and some jobs for which the requirements are so unusual, that they must be treated individually. These exceptions in no way minimize the importance of the general principle. 11-3 6 Classification Should Be Super- vised by Classification Specialists Personnel classification is a technical spe- cialty. The specialists differ from the amateurs in making fewer mistakes and in knowing how many mistakes they make. It may appear easy to examine a man’s record card, to note that he worked as a garage mechanic, and to decide that he would probably be successful as a ma- chinist’s mate. It is more difficult to determine which other qualities are involved in determin- ing whether the garage mechanic will or will not be a good machinist’s mate. And it is also more difficult to determine which men who have not had mechanical experience will be good machinist’s mates. The classification specialist knows how to solve these more difficult prob- lems, and he knows just how successfully they can be solved with present methods. Personnel classification is a technical specialty. It requires technically trained officers to ensure successful classification. 114 DEVICES TO AID MILITARY CLASSIFICATION 11 4 1 Tests of Aptitude and Achievement Aptitude and achievement tests bear a large share of the burden of determining the partic- ular specialty to which each man will be as- signed. Specific aptitude tests have been de- scribed in Chapters 2 to 10. Their construction, standardization, and validation are discussed in Chapters 14 and 15. Achievement tests are discussed in Chapter 17 of Volume 2, Applied Psychology Panel. Interview Aids Several devices to aid the interviewer in se- curing relevant information and to help in assigning appropriate weight to the various items of information were developed. A Point-Score Method for Evaluating Naval Personnel Interview forms listing specific elements of experience considered important for success on destroyers were prepared. A weight was as- signed to each item in terms of its importance as judged by a number of individuals well ac- quainted with the duties to be performed. The total score for any individual could then be determined by summing the weights of the items. The weighting of interview information was by this means made more standardized.7 Interviewing Guides Another aid to the standardization of inter- viewing procedures was the development of interviewing guides for destroyer billets.5’12 Work-Readiness Tests Work was begun in 1945 on the development of short series of questions patterned after the U. S. Employment Service oral trade questions. They were intended for use in determining a man’s readiness (in terms of knowledge) to assume the duties of a particular billet. Only one, for distilling plant operators, was actually constructed.9-10 RESTRICTED 108 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION 11 1 J Shipboard Classification Devices In processing men through shore-based clas- sification centers a great deal of information is obtained about each man’s experience, ability, and potential usefulness aboard ship. There is still, however, a problem of making that in- formation readily available aboard ship. Emer- gencies arise and casualty replacements must be made from within the ship’s company. In order to make these replacements quickly and effectively it is desirable to have some method of determining which members of the crew pos- sess the needed skills or experience. The most efficient solution of this problem consists of having the pertinent data concern- ing each man entered on a single record card. Various companies, the Findex and the McBee for example, have developed simple mechanical methods of sorting these cards so that the cards for men who possess the desired characteristics can be quickly separated from the rest. These systems possess the distinct advantage of making all the recorded information about a man immediately available. There is no for- getting. On three occasions during World War II the Applied Psychology Panel was asked to help in devising mechanical sorting systems for ship- board use. Each of the three systems is briefly described below, Findex Equipment for the USS New Jersey Following the classification of the crew of the USS New Jersey (Section 11.2.1), Findex equipment was installed aboard ship to aid in the rapid location of men needed in the oper- ation, particularly in the emergency operation, of the ship.1 The manual contained detailed in- structions to supplement those furnished by the Findex Company, Sorting the Q-Card Members of the staffs of Applied Psychology Panel Project N-106 and of the Classification and Selection Section, Bureau of Naval Per- sonnel, cooperated in devising a sorting scheme to increase the usefulness of the U. S. Navy Enlisted Personnel Qualifications Card (Q- card). The scheme made provision for coding seven items of information (Rate, Branch, Oc- cupational Skills, Gunnery Experience, Foreign Language Fluency, Rate Designator, and Talker Qualification) along the two perforated edges of the Qualifications Card, An accompanying manual3 described the details of the coding and illustrated how various types of sorting jobs could be done once the cards had been coded. Using templates, the guide holes on the edge of the card can be punched according to a code system which provides for coding the following seven items of information along the two per- forated edges of the Q-card. 1. Branch (seaman, artificer, etc.—eight cat- egories in all). 2. Rate (100 categories in all). 3. Special Occupational Skills (142 catego- ries) . 4. Gunnery Experience (two categories). 5. Foreign Language Fluency (two catego- ries) . 6. Rate Designator (two categories). 7. Talker Qualification (two categories). The branch code number and punch code num- bers illustrate the coding system: Code and Punch Code Numbers Used in Recording Branch Branch Code Number Branch Code Number Seaman 1 SF and 1 Artificer 2 SF and 2 Engine Room 3 2 and 1 Aviation 4 SF and 4 Special 5 4 and 1 Commissary 6 4 and 2 Steward 7 SF and 7 Specialists 8 7 and 1 A second example shows the punch code num- bers for several job titles of the 142 categories of special occupational skill which can be coded on the Q-card. The scheme preserves the existing format and content of the Q-card. It is comparatively easy to use, and the cost of the four pieces of equipment which it demands is nominal. The system itself does not impose any definite limi- tation on the size of the groups of cards dealt with, but it is believed that it can be used most conveniently with groups of less than 500, Experimental Shipboard Personnel Record System The Enlisted Classification and the Test and Research Section of the Bureau of Naval Per- sonnel, Applied Psychology Panel Project RESTRICTED DEVICES TO AID MILITARY CLASSIFICATION 109 N-106, and the McBee Company prepared a Personnel Record File for installation aboard the USS Catoctin.8 The face of the card contains room for re- cording- 53 items of information descriptive of a man’s aptitudes, experience, and training-. A selection of 30 of these items can be coded along- the edges of the card where they are ac- cessible for sorting purposes. Use of Shipboard Filing Systems The record of each man is reduced to a single card of, say, 8 by 10i/2 inches. These cards are filed alphabetically, making it an easy matter to locate information on the experience, train- ing, and special abilities of any individual man aboard ship. Table of Special Occupational Skill Codes (Alphabetical) Code Punch Code Number No. Job Title Hundreds Tens Units 1 Accountant 0 0 SF-1 22 Accounting Clerk 0 SF-2 SF-2 2 Actor 0 0 SF-2 64 Advertising Manager 0 4-2 SF-4 77-a Aerographer 0 SF-7 SF-7 109 Agent, Purchasing 1 0 7-2 36 Air Conditioning and Re- frigerating Engineer 0 2-1 4-2 3-a Aircraft Metal Worker 0 0 2-1 102-a Aircraft Pilot 1 0 SF-2 Figure 1, Record card for shipboard filing system. The record card used was of the keysort type. It is shown in Figure 1 in front view. The back of the card contains room for a chronological record of the man’s naval training and experi- ence. When a need arises for men of specified qual- ifications, simple mechanical sorting can quickly locate them. As examples, one could pick out all men rated as qualified telephone talkers, all men having a particular occupational skill, all RESTRICTED 110 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION men who have been under fire, all experienced in damage control, or all left-handed men. If combinations of skills are desired, it is possible to select on several traits. One could, for ex- ample, locate in a few minutes every qualified telephone talker who had been under fire and who had had experience in damage control. The usefulness of such a record system de- pends upon the thoroughness with which in- formation is entered on the record card and the care with which the records are kept up to date. If those things are done properly, the possession of such a record system makes it possible for the ship’s officers to locate the men they want when they want them. ii-4.4 Mechanical Weighting of Classi- fication Data The most novel development of the Applied Psychology Panel’s work in the field of classi- fication procedures was the invention of a device called the selectometer which automatically weights and combines a man’s scores (or other measures) on each of the factors considered important in selecting men for a particular job.4 The selectometer was developed by Project N-116, Stanford University, in cooperation with officers of the Operational Training Com- mand, U. S. Navy Pacific Fleet. The basic aim of classification is to select the men best suited for each of a number of posi- tions. This selection is made by considering the qualifications of the individuals available for placement and comparing those qualifications with the requirements of each job. This is not easy. Remembering and weighting, more or less simultaneously, qualities which the classifier believes are related to success in different jobs presents an extremely difficult task. In actual practice, the interviewers will frequently over- look some factors, will be inconsistent in the importance attached to the several factors, and will, in fact, seldom know how they weight the factors which they do consider in making an assignment. One way to introduce uniformity into this procedure is to weight the factors and sum the weighted scores mechanically. There are vari- ous ways to do this part of the classification job. It can be done by hand computation or with Figure 2. The optical selectometer. RESTRICTED DEVICES TO AID MILITARY CLASSIFICATION 111 the aid of punched cards. The selectometer does the same job more rapidly. The Optical Selectometer Operation and Construction. Figure 2 is a photograph of the optical selectometer as it appears ready for operation. When a man is to be classified, data taken from his Q-card are set into the control panel by moving pointers to the proper settings. For example, if he is six feet tall the “Height” pointer is moved to the mark “71-72.” When all the other variables (age, test scores, etc.) are similarly treated, the instrument is considered “set up” for the individual. The extent to which the man is qualified for various jobs is then determined from the “read- ing holes.” Each hole represents one job, and the amount of light streaming through each hole determines the degree of qualification for the corresponding job. The intensity of light is measured by a phototube connected to a microammeter through an amplifying circuit. The meter reading thus provides a quantitative measure of fitness for each position. The instrument uses an optical-mechanical method of weighting qualifications. The extent of qualification for each job is indicated by the intensity of the beam of light from a fluorescent light source through the corresponding hole in the front panel. The beam of light is reduced in intensity by interposing filters in proportion to the lack of qualifications which an individual possesses. Thus, the more qualified a man is for a job, the more intense is the beam of light through that job hole. The less qualified he is, the greater is the reduction by means of filters, and when an individual possesses a character- istic which would disqualify him for a job, the use of an opaque filter eliminates all light. The filters are mounted on plates which slide behind the front panel. When a plate is moved to the position which represents a particular score or class interval of a variable, the holes in it are so lined up as to bring the proper filter directly behind the front reading hole. Each of the plates represents one variable. The total amount of light transmitted by the set of filters indicates the degree of qualification for the job. To offer a standardized method of represent- ing the degree of qualification of any category, a set of filters was prepared from 85 mm mo- tion picture film of varying density or light transmission value. For convenience, the filters Table 1 . Filters used for weighting. Filter No. (weight) Desired percentage of light Shade transmission Interpretation for degree of job qualification 0 Clear 100 Best for job, or no adverse 1 Very light 90 relationship Slightly less desirable, or 2 Light 80 probably not quite the best Fairly desirable, or may be 3 Light gray 70 only slight handicap Probably adequate, but like- 4 Medium 60 lihood of some handicap Possible success, but prob- 5 gray Dark gray 50 ably some drawback Doubtful; considerable 6 Dark 40 probability of failure Unlikely to succeed, chances 7 Very dark 20 more definitely against Bare possibility of success; 8 Opaque 0 great risk Reject; very high prob- ability of failure were given numbers from 0 to 8 according to the density. Table 1 indicates the amount of light transmitted by each filter and its inter- pretation in terms of degree of job qualifica- tion. The weights assigned to each category within each of the variables used in making gun sta- tion assignments are given in reference 4, These weights were assigned on the basis of analyses of the duties and requirements of the gunnery jobs for which assignments were to be made, a study of the characteristics of the men avail- able for assignment,2 and conferences with a number of gunnery officers to refine and stand- ardize tentatively assigned weights. The method of weighting used in the se- lectometer combines the advantages of the multiple cutting score technique with those of the multiple correlation method. The selectom- RESTRICTED 112 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION eter, like the cutting score procedure, permits one to eliminate a man on the basis of one dis- qualifying characteristic alone, even though his scores on other variables may be extremely good. On the other hand, when a very high qualifying score has a higher predictive value than a score just above the cutting line, the selectometer, like the regression equation, gives extra weight to high scores; it thus takes full advantage of any predictive differences in the scores above the cutting point. In its application at the Precommissioning Training Center, Treasure Island, California, to classification for destroyer gun stations, the following variables were incorporated in the instrument; GCT, Arithmetic, Mechanical Knowledge, Height, Near Vision, Far Vision, Age, Years of Education, Interest (for a type of gun), Talker Test, Build, Sports in Which Qualified, and Leadership. In its use at other centers some different variables were used ac- cording to the measures which were available in the classification office. The list of variables used in the selectometer does not include all the factors which are im- portant for success in gun stations. For exam- ple, a man’s success in these positions depends upon his motivation, dependability, work habits, willingness to follow orders, ability to get along with others, his morale, his ability to stand up under fire, and a host of other factors. The reason for not including such variables is sim- ply that no adequate measures or estimates of them are obtainable. There is a lack of adequate methods of making such estimates and a lack of personnel and time to obtain such data as might be available. An interesting feature of the selectometer is that the unsatisfactory and unreliable appraisal of some traits becomes immediately apparent when their inclusion in the list of selectometer variables is considered. For example, inter- viewers who think they have been considering a man’s “leadership” in recommending his best assignment find themselves facing a new prob- lem when they are asked exactly how much weight should be given to what degrees of leadership. When more reliable methods of esti- mating such traits become available and when agreement on their importance is reached, they may readily be incorporated into the selectom- eter system. The important variables, rate and experience, were omitted from the selectometer because most officers are already accustomed to using rate and experience in making gun station as- signments. They would be doubly weighted if they were included in the selectometer. Fur- thermore, there is a definite difference in policy among officers with respect to assignment of rated men to gun stations. Reliability of Selectometer Ratings. The re- liability of the selectometer method depends not only upon the consistency of the instrument itself but upon the accuracy of the operators. An experiment to determine the reliability of the method was undertaken with two experi- enced operators. The selectometer raw scores for 101 individuals were obtained on two occa- sions a day or two apart. Product-moment cor- relations were obtained for four typical posi- tions: 5"/38 sight-setter, .84; 20 mm gunner, .96; 40 mm second and third loader, .97; 20 mm trunnion operator, ,90. Analysis of the scores on which there was variation seemed to indi- cate that inaccuracies on the part of the oper- ators constituted the major source of unrelia- bility. From all the present evidence it appears that the consistency of the instrument itself is extremely high. The accuracy of the light filter method was determined by comparing optical selectometer scores with those obtained by the punched card method described below. For 20 mm gun- ners (N = 188) the correlation was .90. When corrections for errors of operators are consid- ered, one may conclude that the method of using light filters for combining weights is highly accurate. The ultimate test of the validity of the selectometer is the overall comparison of se- lectometer scores with gun station proficiency at sea, especially during combat. Practical con- siderations made this test of validity impos- sible. The Electrical Selectometer Figure 3 is a photograph of a demonstration model of the electrical selectometer. In this in- strument four variables are used. They are RESTRICTED DEVICES TO AID MILITARY CLASSIFICATION 113 height, far vision, age, and GCT, as repre- sented by the switches on the front panel. Each stop of the switch represents a category of the variables as indicated in the previous discus- sion. The job switch at the right-hand side of the panel includes a sample of jobs with different requirements. These are 20 mm gunner, 40 mm gun captain, 40 mm pointer, 5"/38 first loader, 5"/38 passer. istic is judged to have for the job. This is, in essence, the opposite of the optical selectometer, where the current is decreased in proportion to the penalty. In the electrical selectometer, a resistor is analogous to the filter in the optical selectometer; a large resistor which permits very little current to pass through is compar- able to a light filter which reduces the light only a small amount, and a very low resistor transmitting a large amount of current has the Figure 3. The electrical selectometer When an individual’s potentialities for vari- ous jobs are to be determined, the variable switches shown in Figure 3 are set for his char- acteristics. Other variables may be incorporated as methods of estimating are developed. The job switch is then turned to each job in succes- sion, and the scores are read directly from the meter. These scores may be used in the same fashion as those obtained from the optical selectometer. In the electrical selectometer, current flow- ing through the meter is increased in propor- tion to the size of penalty which a character- same function as a very dark filter. In other words, in the electrical selectometer, no current through the ammeter means “Fully Qualified,” while maximum current through the meter means “Disqualified.” In order to avoid con- fusion the scale on the meter was reversed, so that 50 means qualified, and 0 disqualified. The wiring of the instrument is shown in Figure 4. For the sake of clarity, only three jobs are shown. When the switches are set for a variable, a conductor of a certain resistance is thrown into the job circuit, thus adding an amount of current proportional to the esti- RESTRICTED 114 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION Figure 4. Wiring diagram for optical selectometer RESTRICTED DEVICES TO AID MILITARY CLASSIFICATION 115 mated degree of disqualification. The source of current is a 12-volt battery, but ordinary 110- volt alternating current through a rectifier tube could replace batteries. An ammeter reading from 0 to 50 milliamperes is used to measure the current flowing in the circuit. By means of a special circuit, as indicated within dotted lines in Figure 4, the distribu- tions may be changed so that the limits for group scores will be more nearly similar for all jobs. The wiring diagram also shows how, by the introduction of another power source, cur- rent may be added to or subtracted from the job circuit so that all the scores on one job may be lowered or raised by any desired con- stant amount. By using the shunt resistance across the meter at the same time, the sensi- tivity of the meter is reduced so that the stand- ard deviation of these scores is lowered. If the distributions were all normal, these circuits would make it possible to obtain exact standard scores directly. The original design of both optical and elec- trical selectometers gave the same weight to a category regardless of the scores on other vari- ables. In certain special instances, it may be found desirable to make the weight assigned to variable A depend upon the value of variable B. In other words, it may be decided that the weights should not be assigned to either vari- able independently but should always be as- signed to a particular combination of scores on the two variables. The method may be extended to deal with special combinations of more than two variables if necessary. The selectometer may, in fact, be constructed to give a certain designated score to any combination of any number of variables.4 This procedure will not only permit the incorporation in the selectom- eter of advantages claimed for the profile comparison method, but it will standardize the procedure in a far more complete fashion than a more subjective method ever allows. Advantages of the Electrical Design over the Optical Design 1. The electrical instrument is easier to con- struct ; there are fewer mechanical problems to be worked out; the circuit is merely a com- pound, not a complex, one. 2. The instrument may be built more com- pactly than the optical model. 3. The use of standardized commercial re- sistors eliminates the problem of having special filters developed. These resistors may be ob- tained commercially with a 5 per cent tolerance or may easily be selected for greater accuracy if so desired. 4. The use of additional switches would per- mit incorporation into the instrument of all the advantages of the “profile” method in any de- sired fashion. 5. The use of circuits to adjust the mean and standard deviation of the distributions would permit direct reading of scores in either abso- lute or relative terms. 6. The electrical instrument is easier to keep in adjustment because of its relative mechani- cal and electrical simplicity. Punched Card Weighting of Selection Variables Since punched cards are in extensive use in many classification centers in the Navy, a pro- cedure was devised to apply this technique to the problems of the kind which stimulated the development of the selectometer. The use of punched cards as described below yields raw scores which correlate highly with scores ob- tained from the present instrument. As indi- cated earlier, the correlation was .90 for a 20 mm gunner. With the punched card method, the same weight numbers may be used as with the opti- cal selectometer. But the numerical values as- signed to these weights are as follows: Weight Numerical number value 0 00 1 03 2 07 3 12 4 18 5 25 6 34 7 54 8 99 For each category of every variable, a group of cards is gang-punched with the numerical value of the weights assigned to this category RESTRICTED 116 PRINCIPLES AND DEVICES FOR MILITARY CLASSIFICATION QUALIFICATIONS CARD DOE, JOHN S 'p &CT. JT VISION 16/20 HEIGHT G4~ ®W!TH DATA FROM QUALIFICATIONS CARD, PREPUNCHED CARDS ARE PULLED OUT FROM PIGEON HOLES. CARDS ARE PREPUNCHED WITH WEIGHTS FOR EACH JOB. (2) MASTER CARD WITH NAME 4 NUMBER (it) CARDS ARE STACKED IN ORDER IN A BOX. HUNDREDS OF SETS MAY BE THUS SELECTED. THESE CARDS ARE THEN SENT TO THE /.AM. OFF/CE. ®THE CARDS ARE TABULATED ON MACHINE NO. 405. BESIDE EACH MANS NAME IS PRINTED THE TOTAL HEIGHT FOR EACH JOB. UNITS DIGIT IS OMITTED FOR SIMPLICITY. SIMULTANEOUSLY THESE TOTALS ARE SUMMARY PUNCHED ON A CARD , ONE PER MAN. (j\ PREPUNCHED \Lucards ape RESOPTED AND RETURNED TO PtGEON HOLE CABINET FOR USE OVER AGAIN. Tjp\ THE REPORT IS SENT TO THE EXECUTIVE OFFICER TO BE USED IN ASSIGNMENT OF MEN TO BATTLE STATIONS /£\ SUMMARY CARDS ARE Vjy INTERPRETED AND USED TOR DISTRIBUTIONS, VALID- ATIONS (FURTHER RESEARCH. Figure 5. Graphic outline of punched card method of weighting selection variables RESTRICTED DEVICES TO AID MILITARY CLASSIFICATION 117 for each job. The same two columns are assigned to each job on all the cards for all categories of all variables so that the sum of the numbers in any set of two columns will give the total score for a job. These prepunched cards are filed in a pigeonhole cabinet, each hole containing the cards for one category. In Figure 5 is presented in graphic form an outline of the general method to be followed in processing an individual through this classify- ing procedure. Evaluation of the Selectometer On the basis of the results obtained with the optical selectometer and of the evidence avail- able from the development of the electrical selectometer and the punched card method, an evaluation of the application of selectometer procedures to classification problems may be made. Whether one uses an optical or an elec- trical model or uses punched cards is a matter of convenience. The principles and advantages are the same. The chief points are: 1. The selectometer focuses attention on the importance of carefully developed job specifica- tions. In order to be incorporated into the selectometer, job specifications must be stated very explicitly. This precise statement facili- tates the validation of billet specifications. 2. The selectometer standardizes the use of billet specifications. By combining and weight- ing scores mechanically, it eliminates any in- dividual differences in the importance attached to various characteristics. Various weighting techniques for combining the variables may be employed. Once the weighting technique has been agreed upon, the selectometer will follow it uniformly for all men. 3. The selectometer provides a method of obtaining qualification scores of each man for a number of different jobs. The time required to do this task for a large number of men and jobs without some partly mechanized system would be prohibitive. 4. Different classification needs can be ac- commodated by revising weights, adding or sub- tracting variables, or raising or lowering cutoff scores. While the selectometer was designed and initially used for making gun station assign- ments it appears to be equally applicable to the problem of making school assignments for re- cruits or, indeed, to help in any classification situation in which a large number of men are to be sorted into a relatively large number of separate groups on the basis of a number of different measures. RESTRICTED Chapter 12 ORGANIZATION OF A CLASSIFICATION PROGRAM FOR RECRUITS Norman Frederiksena Summary The procedures followed during World War II in classifying and assigning enlisted per- sonnel in the Navy are briefly described. Improved classification would result from: 1. Using a primary test battery to separate the men of school quality from those who should be assigned to general detail, and a secondary test battery to determine the particular school to which each man in the first group should be assigned. 2. Improving the interview by standardizing the weighting assigned to various factors on which the interviewer’s recommendation is based. 3. Improving the basis for filling school quotas by establishing assignment pools and by better determination of proper priorities in filling quotas. 4. Continuous research to keep classification tests and methods abreast of changing military requirements. 5. Indoctrinating regular officers in modern methods of personnel classification. 121 OBJECTIVES AND PROBLEMS OF RECRUIT CLASSIFICATION The classification officer at an Army Recep- tion Center or Naval Training Center faces a difficult problem during wartime. Every week thousands of recruits are brought to the center. They represent very nearly the maximum pos- sible range of variation with respect to almost any possible measure except age and physical condition. Some are doctors of philosophy, while others have never attended school; some are geniuses or potential geniuses, while others are feeble-minded; some are highly skilled me- chanics, engineers, and instrument makers, while others are poets, violinists, and artists. Within a period of a few days the classification officer and his staff are required to evaluate the actual or potential abilities of each recruit and assign him to that type of military training or duty where he will make his greatest contribu- tion toward the military prosecution of the war. In peacetime there may be more time avail- able to the classification officer, and the weekly flow of men may be smaller, but the problems of proper classification remain. For the effi- ciency of a military force depends upon the efficiency of the men who compose it. Classify- ing and assigning each man to the job that he can do best will make the whole force an effec- tive one. The objective of military classification is, then, to make the most effective use possible of manpower. This objective may be accomplished by evaluating the aptitudes, abilities, skills, in- terests, educational and occupational experi- ence, and physical and personal characteristics of each man and on the basis of these evalua- tions assigning the man to a particular type of military training or duty. (Examples of the tests used for classification and evidence of their value were given in Chapters 2 to 10.) In addition, the aim is to make the optimum assignment of each man in the minimum amount of time. Given sufficient time, men might eventually “shake down” or gravitate to the billets for which they are best qualified, even without careful classification procedures; but it is important, particularly in wartime, that each man be given a satisfactory assign- ment as quickly as possible. The saving of time in making suitable assignments, especially in the early stages of a war, may in effect in- crease considerably the manpower available for combat. More specifically, the objectives of classifica- tion of recruits are to answer such questions as the following: a This chapter is based chiefly on the experience of NDRC Project N-106 in helping to develop the Navy’s classification tests and procedures. 118 RESTRICTED CLASSIFICATION PROCEDURES IN USE BY THE NAVY 119 1. Who (among the thousands of recruits to be classified in a particular week) already possess sufficient knowledge, skill, and quali- ties of leadership to warrant their being sent immediately to their duties as technicians or noncommissioned officers? 2. Who possess the aptitudes and intellectual qualifications which warrant sending them to a Service school for training in some military specialty? 3. Of those who are judged to have the apti- tudes which warrant specialist Service school training, who shall be trained as radarmen, who as gunners, who as quartermasters, who as tank drivers, who as radio mechanics, etc.? 4. Who lack the qualities which would justify investing time and money in special technical training and therefore should be assigned to general, nontechnical duty? These are samples of the questions which the classification officer must answer. The job of classifying the recruits is made difficult, not only by the fact that the men are coming con- tinuously in tremendous numbers and must be processed within a short period of time, but also by various other considerations. For one thing, a large proportion of the men have, be- cause they are young, no background of civilian occupational experience which would help in making assignments; therefore aptitudes, or ability to acquire specific skills, must be evalu- ated in advance of training. A conscript army in peacetime would be composed of young re- cruits without significant occupational experi- ence. In the case of older men, occupational experience may be a factor which can be used in classification; but even then its usefulness is limited, since civilian occupations do not bear a one-to-one correspondence to military jobs. Few if any civilian jobs are similar in nature to those of a military gunner or torpedoman. Some civilian occupations, such as insurance salesmen and textile workers, are available in numbers far greater than needed by the Serv- ices. Furthermore, assignments must some- times be made to schools where, because of the rapidity of development of new military tech- niques, the aptitudes required and even the curriculum taught are not fully known. In short, large numbers of men are available whose skills and qualifications are not obviously ap- parent; they must be fitted to a great variety of military jobs, some of which require rare or unusual aptitudes, or aptitudes which are diffi- cult to evaluate, or even aptitudes the nature of which is not fully known. 12 2 CLASSIFICATION PROCEDURES IN USE BY THE NAVY The classification procedures currently in used by the Navy5-6 are described briefly in this section. Treatment of Army classification pro- cedures is not included because almost all the work of the Applied Psychology Panel on re- cruit classification was done in the Navy. Testing and Interviewing The recruit is given a physical examination (including psychiatric screening) by the Medi- cal Department soon after his arrival at the station. The report is forwarded to the Classifi- cation Department. In the days which immedi- ately follow, the recruit spends most of his time in activities sponsored by the Selection Depart- ment: (1) he takes the tests of the U. S. Navy Basic Classification Test Battery (see Chapter 2) ; (2) he listens to lectures and sees movies calculated to inform him about the schools and rates available to enlisted men; and (8) he fills out an “Aid to the Interview Blank” on which he describes his educational and occupational history and gives other information which the classification interviewer may later use in mak- ing his classification. Preparation of Record Forms In the meantime, the preparation of two important personnel record forms is begun. The first of these is the Classification Depart- ment Electric Accounting Punched Card, In it is punched information which the Personnel Department supplies—Service number, name, rate and grade, company number, etc. As soon as the answer sheets for the Basic Battery tests have been scored (by means of the test-scoring machine), the scores are also punched in the card. The second personnel record is the Enlisted RESTRICTED 120 ORGANIZATION OF A CLASSIFICATION PROGRAM FOR RECRUITS Personnel Qualifications Card (Q-card), which, when processing is complete, becomes a part of the recruit’s permanent Service record which accompanies him throughout his service in the Navy. The Q-card is printed on continuous form stock for use with an electric accounting ma- chine. By using this machine, the data so far punched in the Electric Accounting Card are printed on the top two lines of the Q-card. Then additional data from the Medical Depart- ment report and the names of particular schools or duties for which the recruit is quali- fied or disqualified are entered manually. The Interview The recruit is now ready for the interview, which is conducted by a Specialist (C). Spe- cialists (C) are petty officers who have been selected because of special qualifications (such as experience with the U. S. Employment Serv- ice) to serve as classification interviewers. They are given a special course of training prior to classification work, as well as in-Serv- ice training. The interviewer has at hand the Aid to the Interview Blank as previously filled out by the recruit, and the Q-card which is partially processed as described above. During the interview the Specialist (C) adds to the Q-card additional information pertaining to educational and occupational history, hobbies, and the like. 12.2.2 Recommending Assignments Now comes the most important function of the Specialist (C). He makes recommendations, on the basis of all the information so far ob- tained, as to the assignment of the recruit. The success or failure of the classification procedure hinges largely on the wisdom of these recom- mendations. The interviewer ordinarily makes two recom- mendations on the basis of the recruit’s test scores, his occupational experienced physical qualifications, interests, etc. The first recom- mendation represents the most appropriate assignment, in the opinion of the Specialist (C), and the second recommendation repre- sents the next most appropriate assignment. The recommendations may be for assignment to general detail, to an elementary Service school, or to an operational school. In addition to making specific recommenda- tions, a Quality Classification Code is assigned to each recruit by the Specialist (C) ; this code is supposed to be a qualitative rating in terms of the recommended assignment. Code 1 is used to designate men judged to be exceptionally well qualified for the recommended school as- signment. Code 2 men are, in general, as well qualified as Code 1 men but lack a “definite pat- tern of qualification” and therefore can be as- signed to a school similar to that of the first recommendation. Code 3 men do not have the high qualifications of the above groups for the recommended assignment but are sufficiently well qualified for assignment if needed. In addi- tion, codes may be assigned which indicate recommendations that the man be rated, that he be held for reassignment, or that he be re- interviewed, perhaps by a psychiatrist. The classification of the man may also be postponed until special tests can be given; this is done in the case of men tentatively selected for a cer- tain school on the basis of Q-card data. Code 0 indicates lack of qualifications for school assignment; this quality classification always accompanies a recommendation of assignment to general detail. The recommendations and Quality Classifi- cation Code are recorded by the Specialist (C) on the Q-card, and later punched into the Elec- tric Accounting Card. These punched cards are now ready for sorting. 12-2 3 Filling Quotas As soon as possible, a summary of the recom- mendations is sent by teletype to the Enlisted Distribution Division of the Bureau of Naval Personnel. At the Bureau, quotas are made up, on the basis of fleet demands and the number of men recommended for Service school at each of b A booklet called United States Navy Occupational Conversion Tables (October 17, 1942) has been pre- pared for the purpose of aiding the interviewer in evaluating civilian work experience. This booklet gives the Navy rates which are related to a large number of the job titles in the Dictionary of Occupational Titles of the U. S. Employment Service. RESTRICTED SUGGESTIONS FOR IMPROVEMENT OF THE NAVY’S CLASSIFICATION PROCEDURES 121 the various naval training centers. Then the Bureau sends to the Selection Department its quotas—the number of men who must be sent to each of the various duties and training activ- ities. The Order of Assignment It is, of course, not likely that for a given week the distribution of Service school quotas will match the distribution of recommended school assignments. A specified order of assignment which may be applied in making assignments to any Service school has there- fore been developed. In filling quotas, the cards are sorted and men selected according to the following order: 1. 1st recommendation, quality class 1 2. 1st recommendation, quality class 2 3. 2nd recommendation, quality class 1 4. 2nd recommendation, quality class 2 5. 1st recommendation group, quality class 1 6. 1st recommendation group, quality class 2 7. 2nd recommendation group, quality class 1 8. 2nd recommendation group, quality class 2 9. 1st recommendation, quality class 3 10. 2nd recommendation, quality class 3 11. 1st recommendation group, quality class 3 12. 2nd recommendation group, quality class 3 If a quota for quartermaster school, for ex- ample, were to be filled, the cards would be arranged according to the order of assignment, and the quality class “1” men whose first recommendation is quartermaster would be chosen. If sufficient cards are found to fill the quota, the job is finished so far as quarter- master school is concerned, and one can go on to another school’s quota. But if there are not enough quality class “1” men whose first recommendation is quartermaster, it is neces- sary to make further selections; quality class “2” men recommended for quartermaster school would next be picked out, and so on. By “first recommendation group” is meant a group of schools with selection requirements similar to those of the school recommended first; for example, if the quartermaster quota had not been filled after choosing men in the first four categories of the order of assignment, the next step would be to sort out quality class “1” men recommended for schools in the same group as quartermaster—aerographer’s mate or Specialist (Y) control tower operator schools. Completing the Personnel Records After the cards have been sorted to fill all quotas, the selection lists are printed from the cards by means of the electric accounting ma- chine. The assignment and date of transfer are punched in the cards, which are then dupli- cated. One card is filed by the Classification De- partment and the duplicate goes with the man to his Service school assignment; these dupli- cate cards are later used for research purposes. The Q-card is photographed and then inserted in the man’s service record; the film is sent to the Navy Department. 12 3 SUGGESTIONS FOR IMPROVE- MENT OF THE NAVY’S CLASSI- FICATION PROCEDURES A good deal of the research4 on recruit clas- sification has been based on the punched cards which were mentioned above. This research, to- gether with firsthand observation of procedures and talks with classification officers, has led to a number of suggestions for improvement of the Navy classification procedures. Some of the more general suggestions are discussed in the following section. 12-31 Use of Primary and Secondary Test Batteries The first suggestion for improvement of the classification program is based on the diffi- culties that are inherent in using one basic bat- tery of aptitude tests for all the various pur- poses involved in selection. Use of the same tests for such diverse purposes as separating school from nonschool material and separating RESTRICTED 122 ORGANIZATION OF A CLASSIFICATION PROGRAM FOR RECRUITS candidates for fire control school from candi- dates for torpedo school is likely to result in tests which are not entirely suitable for either purpose. Furthermore, administration of the entire battery to the large group of men who will not go to Service school is wasteful of test- ing and test-scoring time. It is possible to administer a greater variety of tests without increasing the total time spent in test administration by the device of using a short primary battery of tests for the purpose of selecting the school material, and using sec- ondary test batteries for the purpose of assign- ing men to specific Service schools. The details of such a program as is here suggested would of course have to be modified in accordance with research findings; the fol- lowing suggestions are tentative. The Primary Test Battery The primary battery, which would be ad- ministered to all recruits, might consist of only two tests: a purely verbal test, somewhat like the present General Classification Test, and a mechanical test, which might consist of items resembling the pictorial items of the present Mechanical Knowledge Test. This second test might be entirely a picture test, in order to lower its correlation with the verbal test as much as possible. Item analysis techniques should be used, in developing these tests, to ensure as low a correlation as possible between the verbal and the mechanical tests. This primary battery of two tests could then be used to separate those men who are unlikely to succeed in any type of Service school from those who are likely to succeed in some Service school. (The cutting scores, selection ratios, or other devices would of course have to be de- termined on the basis of research investiga- tions.) The primary battery would be useful, not only for identifying the men who are the best risks so far as Service school success is con- cerned, but also to classify these men into groups depending upon the general type of Service school for which they should be con- sidered. The scheme might work somewhat as follows: men low on both verbal and mechanical tests should be assigned to general detail; men who are high on verbal but low on mechanical should be considered as candidates for a “clerical” type school (yeoman, signalman, radioman, etc.) ; men who are high on mechani- cal but low on verbal should be considered as candidates for a “mechanical” type of school (gunners mate, basic engineering, electrical, metalsmith, etc.) ; and men who are high in both should be considered as candidates for the more highly technical schools (fire control, radio materiel, etc.). The Secondary Test Batteries The next step in classification would be to administer to the men in each group a sec- ondary battery of tests especially designed to aid in assignments to schools within that group; for example, the men in the clerical groups might be given a battery of tests in- cluding a clerical test, a spelling test, a radio code aptitude test, and a blinker aptitude test. The men in other groups would similarly be given batteries of tests appropriately selected to aid in assignment to the schools of the group. Administering a primary and a secondary test battery would take no longer than the present method of administering every test to every man. Some men would take only the primary battery; the others would, in general, take no more tests than they do at present. This system would probably result in more valid tests, since each could be designed for a specific purpose. It also admits the possibility of greater use of individual tests, because, by subdividing the number of men to whom a secondary test battery must be given is considerably reduced. The objection to the procedure is that problems of scheduling are involved. It would no longer be possible to order an entire company to take a single battery of tests. Methods would have to be evolved for assigning men to take tests individually instead of by companies. 12.3.2 Improvement of the Interview A second suggestion for the improvement of the classification procedure arises from cer- RESTRICTED SUGGESTIONS FOR IMPROVEMENT OF THE NAVY’S CLASSIFICATION PROCEDURES 123 tain researches which seem to indicate that as- signment to Service schools on the basis of test scores alone would result in somewhat more successful classification than assignment on the basis of interviewers’ recommendations.2’ 3 The interviewers, it seems, tend to weight too heav- ily the information they themselves obtain, at the expense of the aptitude test scores. Fairly rigid requirements with respect to use of test scores should, accordingly, be set up, and the interviewers should be required to follow these standards. Furthermore, these studies suggest the importance of further study of the validity of the other items of information on the Q-card, and if the results justify it there should be set up specific rules for weighting each item of information in making the school recommenda- tion. A standardized procedure for evaluating the various factors that are significant in mak- ing an assignment was developed by one of the Applied Psychology Panel projects.1 The pro- cedure, called a point-score method, involves the assignment of uniform weights to carefully selected and explicitly stated factors. Evalua- tions based on such a procedure would certainly be far more reliable and objective than the ordinary interview evaluation. The validity of the procedure would depend upon the care with which the factors and their weights are se- lected. The objection to the method is that the factors and weightings would vary from one assignment to another and points would need to be assigned separately for every assignment considered. Improvement of Procedures Involved in Filling Quotas The third suggestion pertains to the quota system of allocating recruits to various activi- ties. As was described above, the fact that quota distributions fail to match the recom- mended school assignment distribution has necessitated the use of a complex Order of Assignment by which the sortings progres- sively take men less and less well qualified for the assignment. In practice, not only were all twelve categories of the order of assignment used, but it was sometimes necessary to go beyond this order and take men recommended for general detail or for schools entirely outside the Order of Assignment categories.2 Improving Procedures for Determining Quotas The situation could be vastly improved by two courses of action. The first and most funda- mental of these steps is to improve the pro- cedures for setting up the quotas themselves. (This is, of course, not a responsibility of the Classification Department.) In the past it has sometimes appeared that quotas were deter- mined in a somewhat whimsical fashion; the quota for a given school might suddenly double or triple in size, while a large quota for another school might suddenly disappear entirely. More adequate canvassing to determine fleet needs well in advance, together with careful study of the summaries of recommendations supplied by local selection officers, would go a long way toward making it possible to assign most men to the training for which they have been more specifically recommended. The problem of priorities in filling quotas needs careful consideration. At present little or no data are available on the quality of men needed for various types of duty; yet definite priorities have in effect been established. The “Eddy Test” for pre-radio materiel school candidates is administered at the time of in- duction, thus making it possible to make an early selection of only very high quality men for these schools. Fairly high cutting scores on Basic Battery tests have been set for all enlisted personnel given air training assign- ments. Men for the amphibious forces and armed guard, on the other hand, are assigned after the best men have been chosen for other schools, and those left tend to be of rather low caliber. This establishment of priorities in classification no doubt has some justification; but no data as yet exist which indicate whether or not men of high quality are needed for radio materiel more than, for example, fire control or that they are needed more by the air arm than for surface ships. It would seem that every branch of the Service needs at least some RESTRICTED 124 ORGANIZATION OF A CLASSIFICATION PROGRAM FOR RECRUITS men of high quality; although the proportion may vary, competent men are needed to pro- vide leadership in each branch. Research is needed to answer such questions and eventually to determine more precisely how men with various degrees of competence should be dis- tributed. Establishing an Assignment Pool A second course of action which would help in ensuring the assignment of more men to their recommended duty would be the establish- ment of a pool of men whose classification has been completed and who are awaiting assign- ment. The present practice at most centers is to assign the men classified each week on the basis of that week’s quota demands. Delaying assignment only one week would go a long way toward making it possible to compensate for fluctuations in quotas. Such a pool was estab- lished in an unofficial way at one center, with good results; it would have been done at other centers but for the lack of space in which to house the men. With the existence of a pool from which to draw and with quotas which are made up with due consideration of long-term fleet needs and available men, it should be pos- sible to select all men on the basis of the first half of the Order of Assignment and thus to place a much larger proportion of the men in duties for which they are recommended. A pool of men from which quotas can be filled could easily be established merely by making men available for assignment any time during, say, the last three weeks of basic training. The flexibility needed for filling quotas would thus be accomplished without the necessity of addi- tional housing facilities. The time of the men in the “pool” would be occupied by training while awaiting assignment. The only disad- vantage is that the period of basic training might be shortened somewhat for some of the men. The Need for Research The need for research on classification pro- cedures has been mentioned several times in the foregoing paragraphs. In order to keep selec- tion procedures abreast of changes in methods and techniques of warfare, a permanent organi- zation for conducting a continuing program of research is needed. 12 3 5 Indoctrination of Officers In order for the values of a good classifica- tion procedure to be fully utilized and to be of maximum service in efficient utilization of military manpower, it is necessary that officers outside as well as in the Classification Depart- ment have some understanding of classification procedures and sympathy with the objectives of classification. The consequence of lack of un- derstanding is revealed by an extreme case where the classification of a large group of men had been completed and the assignments re- ported to the Personnel Department. The Per- sonnel Department, however, ignored the assignment lists and made their own assign- ments on the purely arbitrary basis of alpha- betization of names. The indoctrination of such a large and varied group of men as the officers in the United States Navy is, of course, a diffi- cult task, especially in the hurry of wartime training. A few weeks’ training in classifica- tion and personnel procedures might be of very little value when given along with navigation, gunnery, communications, and a host of other important subjects in an indoctrination school. In peacetime, however, it would be possible to train the majority of the future officers of the Navy who would become the key men in the event of another war. This training can and should be started at the U. S. Naval Academy at Annapolis. RESTRICTED Chapter 13 ORGANIZATION OF AN ADVANCED CLASSIFICATION PROGRAM Norman Frederiksen a Summary The methods used by the Navy in advanced classification and reclassification of enlisted personnel are briefly described. Several Applied Psychology Panel projects have contributed procedures or recommenda- tions for the improvement of advanced classifi- cation. The principal ones are: 1. More extensive construction and use of objective methods of determining the actual proficiency with which a man can perform the duties of a particular billet. A number of ex- amples of performance tests developed for this purpose are described. 2. Improvement in the methods of rating shipboard proficiency. There are situations in which performance tests cannot be used but in which ratings by superiors can be obtained. Methods of obtaining more reliable ratings are given, 3. Use of brief oral tests, called work readi- ness tests, which can be administered by a clas- sification interviewer to determine how much a man knows about the details of a particular type of duty. 4. The indoctrination of line officers in the methods and information already available which they could use in making shipboard re- assignments. 5. Improved methods for coding and filing the personnel data available for the men as- signed to a particular ship or unit make those data more easily available and more likely to be used. 13 1 OBJECTIVES AND PROBLEMS OF ADVANCED CLASSIFICATION The objectives, problems, and techniques of advanced classification are in general similar to those of recruit classification, which was discussed in the preceding chapter. Both are concerned with the assignment of men, on the basis of actual or potential abilities, to those types of military training or duty where they can make their greatest contribution to the winning of a war. Both are concerned with achieving the optimum assignment for each man in the minimum amount of time. The differences, however, are sufficiently great to justify their separate discussion. The procedures involved in the classification of re- cruits are concerned primarily with the evalu- ation of a man’s potentialities for training, since the majority of recruits have had no training or experience directly related to Navy jobs; advanced classification procedures, on the other hand, more often involve the evaluation of training and experience in the Navy with respect to a specific duty or billet. Assignments in recruit classification are based upon infor- mation rather hurriedly obtained through tech- niques of aptitude testing and interviewing, while advanced classification assignments can be based to a greater extent on actual profi- ciency in Navy jobs, as reflected in records of performance in a training situation, in a billet aboard ship, or on tests of proficiency. Ad- vanced classification, furthermore, should in- volve the balancing of crews so that all ships and departments have their fair share of su- perior men as well as those of average or in- ferior ability. Advanced classification is at once more com- plex and more simple than recruit classifica- tion. It is more complex in that the specific assignments that can be made are more varied and depend upon a greater variety of measures. They are simpler in the fact that assignments can be based more often on measures of pro- ficiency in a given task, rather than on predic- tions of success from measures of aptitude. Advanced classification may be thought of as including any classification procedures that a This chapter summarizes the contributions and rec- ommendations of a number of Applied Psychology Panel projects. RESTRICTED 125 126 ORGANIZATION OF AN ADVANCED CLASSIFICATION PROGRAM occur other than the original assignment of the recruit to some duty or training. Since a man may be reclassified a number of times during his career in the Navy, advanced classification is seen to be an activity which pervades all parts of the Navy, both ashore and afloat, and which is of even greater magnitude than the task of recruit classification. Some typical questions which might be answered by ad- vanced classification are as follows: 1. Who among the members of the crew of a DD are qualified to use sound-powered tele- phones ? 2. What battle station should be assigned to each crew member? 3. Who from a pool of available men should be assigned to each billet on a ship which is about to be commissioned? 4. What men should be recommended for striker billets? 5. What transfers between ships should be effected in order to obtain equitable distribu- tion of abilities? 6. What men need special operational or advanced training to prepare them for assign- ment to the departments, divisions, or battle stations aboard ship for which they are recom- mended? 7. What men should be assigned to ships’ company billets at shore stations? 8. What men need refresher training to keep them up to high standards of proficiency in their billets? 9. What men should be selected for officer candidate training (such as the V-12 pro- gram) ? 10. What members of the crew have useful occupational skills or special abilities (inter- preters, speedboat coxswains, cartographers, mechanical draftsmen, barbers, movie oper- ators, entertainers) which are not indicated by rate, and how can these men be picked out when needed? 11. What men meet the minimum standards of visual acuity and night vision for duty as night lookouts? These questions are representative of those which must be answered in advanced classifi- cation activities. The procedures currently in use by the Navy for supplying the answers to such questions are described in the following section. 132 ADVANCED CLASSIFICATION PROCEDURES IN THE NAVY21 Advanced classification centers are located at receiving stations, training stations, Navy yards, operating bases, and other centers where personnel may be assigned to specific duties. The staff of the classification center processes all men who pass through the station to which the classification center is attached, and also, in some cases, men attached to other nearby shore stations or ships. The processing includes the administration of tests, interviewing, bringing the Enlisted Personnel Qualifications Cards up to date, and making specific recom- mendations as to the assignment of each man. 13 21 Test Administration The general policy with respect to test ad- ministration is to make sure that the record of test scores on the Enlisted Personnel Qualifi- cations Card (Q-card) is complete. If Basic Battery test scores are already recorded on a man’s card, the tests are not readministered (unless there is evidence that the scores are in error). According to directive, if any of the scores are lacking, the appropriate tests should be given and the scores entered. In addition to Basic Battery tests, various other group tests may be given when appropriate, for instance the Radio Technician Selection Test, the Winchman and Hatchman Test, Oral Trade Questions, and the Literacy Test. A number of special apparatus tests are also available, such as the Sonar Pitch-Memory Test, the Telephone Talker Test, and various visual tests for meas- uring such variables as acuity, phoria, depth perception, and color discrimination. 13.2.2 Interviewing and Recommending Assignments One important function of the interviewer is to bring up to date the Q-card, adding any missing information which should have been recorded when the card was first prepared and RESTRICTED SUGGESTIONS FOR IMPROVEMENT OF ADVANCED CLASSIFICATION 127 filling in the more recent history of naval ex- perience and training. Cards which are obso- lete or inaccurate are replaced. On the basis of the Q-card information, including test scores and records of experience and training, the interviewer then recommends assignment of the man to some duty or billet. The interview- er’s general evaluation of the man’s perform- ance is also taken into account. The type of assignment will depend in part on the particu- lar situation—if a man is a member of a ship’s crew which is being processed, the recommen- dations would be in terms of that ship’s watch, quarter, and station bill. The recommendation is made by matching the qualifications pos- sessed by the man to the requirements of a particular billet; this can be done well only if the interviewer is well acquainted with the requirements for each billet. For example, he must know whether or not use of sound-pow- ered phones is involved in a particular assign- ment ; if so, the man assigned to that billet must be rated as a qualified telephone talker as well as meet the other specifications for the billet. In addition to recommending billet assign- ments for each man who is processed, the ad- vanced classification center may perform other services. Recommendations may be made for striker or petty officer billets; a recommended battle bill may be prepared in which specific general quarters stations are recommended; cross index files of potential strikers may be prepared (arranged according to both names and shipboard ratings) ; and cross-index files of special skills may be developed. 13.2.3 Indoctrination of Ship Personnel The classification center furthermore at- tempts to make provision for the continuation of the classification program aboard ship, through the indoctrination of ships’ officers as well as by providing cross index files for use aboard ship. In particular, one officer from each ship, designated by the commanding officer to serve as shipboard classification officer, is given special training in the use of the avail- able data for selecting strikers, assigning bil- lets and battle stations, and effecting transfers among divisions. A shipboard classification yeoman is also trained to process Q-cards, ad- minister and score tests, and keep up-to-date cross index files and files of Q-cards. The above is a brief account of advanced classification procedures. The various advanced classification centers actually differ somewhat in their mode of operation, because of local variations in equipment, personnel, and prob- lems. Some are more effective than others be- cause of better cooperation with other related departments. Some have been particularly suc- cessful in furthering the benefits of classifica- tion through indoctrination of shipboard offi- cers, while others have failed to take steps to ensure the continuation of the program aboard ship. 133 SUGGESTIONS FOR IMPROVE- MENT OF ADVANCED CLASSIFICATION The suggestions for improvement of ad- vanced classification procedures which are con- tained in this section are not new ideas; in many cases, the officers in charge of classifi- cation are fully aware of the problems and have already taken appropriate steps to correct the undesirable situations. However, the Applied Psychology Panel has made contribu- tions in cooperation with classification officers; some of these contributions are discussed below. 13,31 Use of Proficiency Measures As was stated earlier, advanced classification has the advantage, which recruit classification does not, that the men to be assigned have been in the Navy for some period of time; it is therefore possible to make use of evaluations of their Navy experience in making assign- ments. It is of some value merely to know, for example, that a man has graduated from a class A gunners mate school and served as first loader of a 40 mm gun; but it would be con- siderably more important to know in addition how well he could carry out the duties of a gunners mate—how frequently he caused jams through improper loading, how adept he was at RESTRICTED 128 ORGANIZATION OB" AN ADVANCED CLASSIFICATION PROGRAM casualty analysis, and how quickly he could re- place broken extractors. If he is exceptionally proficient, he might be recommended as gun captain; but if, as is entirely possible, he is shown to be inept, he might be recommended for some quite different type of duty. Optimal assignments are most likely to be made when adequate measures of proficiency are available. Prediction of success in a given activity based on knowledge of past success in that activity is more likely to be correct than is prediction based on aptitude tests or on rec- ords which show merely the amount of relevant experience. Advanced classification has an ad- vantage over recruit classification in that the men to be assigned have been in the Navy for some time. It is therefore possible to use evalu- ations of their Navy experience in making assignments. The first suggestion for improv- ing advanced classification is that greater use be made of measures of proficiency in Naval duties. Performance Tests In the case of certain Navy rates, proficiency in the rate can be measured fairly adequately by means of one or two performance tests; often one test is adequate to measure profi- ciency in that aspect of the duty which is cru- cial, i.e., which is the most common reason for ineptitude. In some cases these tests are of such a nature that they could be administered at advanced classification centers; in other cases they could most efficiently be given at operational or advanced training activities. The following paragraphs describe several illustra- tive performance tests. Radio Code Receiving Tests. A set of radio code receiving tests recorded on phonograph records, which was developed by the Applied Psychology Panel, is suitable for measuring the accuracy with which operators can copy plain language and message-type (coded) material sent at approximately the same speed as is used for U. S. Navy “Fox Schedule” broadcasts.19 This test would be very useful at a classifica- tion center for determining the relative profi- ciency of radiomen and radiomen strikers and recommending their assignment to particular billets. A certain amount of special equipment (phonograph turntable with pickup and ampli- fier, headphones, and telegraphic typewriters) is needed, but the test could easily be admin- istered and scored by a Specialist (C). A previ- ous set of tests of a similar nature5 was used in one Operational Training Command to assist in assigning balanced crews of radiomen to ships. Signalman Tests. It would be possible to de- velop tests for signalmen which could be used for similar purposes. Flag hoist spotting tests have already been developed by the Standards and Curriculum Section of the Bureau of Naval Personnel; these consist of colored slides of flag hoists which can be projected, and the men can be required to record the messages. This test and other signaling tests are so far used only in class A signalman schools but probably could be adapted for use at advanced classifi- cation centers. Telephone Talker Test. The only perform- ance test of proficiency which has been used extensively in advanced classification proce- dures is the Telephone Talker Test,1 which was prepared by the Applied Psychology Panel. Specially trained men are used in administer- ing this test; the test is given by conducting an interview over sound-powered phones. The testee is required, during the interview, to re- peat Naval commands commonly used aboard ship, to read a paragraph containing all the phonetic sounds in American speech, and to speak extemporaneously. He is rated on loud- ness, rate, articulation, etc.; on listening abil- ity; and on memory span. Finally a rating of “well qualified,” “qualified,” or “not qualified” is given. The coordination of various depart- ments of a ship through transmitting orders by sound-powered telephone accurately and understandably is obviously of great impor- tance ; the elimination of poor telephone talkers from billets which require the use of telephonic communication is essential to the safety of a ship and its success in combat. Gunnery Proficiency Tests. Some types of proficiency tests require gear and personnel which would make testing impractical as part of the activities of the classification center it- self. This need not, however, prevent the admin- istration and use of such measures. The tests RESTRICTED SUGGESTIONS FOR IMPROVEMENT OF ADVANCED CLASSIFICATION 129 can be given at the termination of the training period in an operational or class A school, and, if the results are entered in the men’s service records, they can be employed just as success- fully as if given at the classification center by classification personnel. For example, tests have been developed for measuring proficiency in each of the three jobs involved in operation of the 20 mm gun—the jobs of gunner, loader, and sightsetter.b The battery consists of four tests—Dry Tracking, Tracer Position Estimation, Range Estima- tion, and Dry Loading. Men in destroyer pools are sent to antiaircraft training centers for training in operational gunnery. It would be quite feasible to give the 20 mm gunnery pro- ficiency test battery at the termination of this training and to send the results to the classifi- cation center which is attached to the destroyer pool. The classification center could then rec- ommend assignment to gunnery billets on the basis of measured proficiency, rather than merely on such variables as Basic Battery test scores, height, weight, vision, civilian experi- ence, and interests. It is quite feasible to de- velop and use in a similar manner proficiency tests in operational gunnery for the larger guns for director personnel. Achievement Testing at a Landing Craft School. A situation which is particularly apt for illustrating the possibility of improving classification by use of proficiency measures was found at a school for training crews of landing craft.18 At this school the crews are trained for Landing Craft, Vehicle-Personnel (LCVP’s). The crew of an LCVP consists of a coxswain, a signalman, a deckhand, and an en- gineer. (The engineer’s billet requires special- ized training, not given at this school; so the selection of engineers will not be considered here.) All the men (except engineers) are given essentially the same training for the first four weeks, since it is desirable that any man be able to take over the duty of any other, in event of a casualty. After the first four weeks, training becomes more specialized, with men designated as coxswains spending relatively more time in boat operation, signalmen in learning semaphore, blinker, etc., and deck- hands in learning seamanship. The classification procedure in use was to select men for each billet prior to any training; classification was based on age, physique, vision, previous experience, and Basic Battery test scores. This procedure was used in spite of the fact that the situation was made-to-order for selection on the basis of achievement during the first four weeks of training. The recom- mendation was made by the Panel project, after certain achievement tests were intro- duced, that signalmen be selected after the pre- liminary training which is common to all men, and that the selection be based in large part on signaling proficiency as shown by achievement test scores. A similar procedure could be used for selecting coxswains, if suitable measures of proficiency in maneuvering an LCVP after four weeks of training could be developed. The men not chosen for either coxswain or signal- man billets would become deckhands, which is the least specialized of the LCVP assignments. Achievement Testing in Other Schools. The use of grades in class A or P schools and in advanced schools as a factor in selection is a possibility which has not been seriously con- sidered in advanced classification procedures. Unless the school grades are more carefully de- termined than is usually the case, it is entirely proper not to consider them. However, steps have been taken by Panel projects, working in cooperation with Naval officers, to improve Service school grades, particularly by supply- ing tests, both of the performance and paper- pencil variety. Given school grades whose mean- ing is standardized and which are based on test situations resembling the shipboard working situations, it would seem that classification might be improved through consideration of school grade. Studies of the validity of classification pro- cedures by correlating the measures used in classification with some measures of success in shipboard duties have rarely been made, mainly because of the difficulty of achieving a suitable criterion of performance in the fleet. One study,10 however, showed that grades made b These tests were developed by Project N-106 of the Applied Psychology Panel; the work is being continued by the College Entrance Examination Board under a contract with the Navy Department. RESTRICTED 130 ORGANIZATION OF AN ADVANCED CLASSIFICATION PROGRAM in submarine training courses are of some value in predicting ratings by submarine offi- cers of the performance of enlisted men aboard ship—of greater value, in fact, than any of the other predictive factors studied. This was in spite of the fact that standardized tests of proficiency were not used in determining school grades. In a number of the Navy’s class A schools a definite attempt has been made to improve school grades—by issuing directives describing in detail what is to be included in school grades and, more important, by devising tests by which knowledge and skills can be accurately eval- uated.2- 3*9-12> 14>17’19' Also in the operational schools of the Amphibious Training Command, Atlantic Fleet, definite progress has been made in developing school grades of increased relia- bility and validity.11 The use of school grades which are based on well-designed performance tests and written examinations should be of considerable value in making assignments; pro- vision should be made for standardizing grad- ing systems, reporting the grades to advanced classification centers, and using them along with other measures in making assignments. If certain of the performance tests on specific pieces of gear (such as disassembly-assembly of a torpedo main engine) are of sufficiently high reliability, it might be well to report their scores separately. The more nearly the meas- ures of proficiency can be made to correlate with specific shipboard assignments, the greater their value to classification centers. Ratings of Shipboard Proficiency The measures of proficiency so far discussed have been of the sort that could most efficiently be administered in connection with a training program. Many of the men to be processed in an advanced classification program come di- rectly from the fleet, and considerable time may have elapsed since the termination of training and hence the determination of pro- ficiency by means of performance tests. It is highly important that proficiency measures used for making assignments be recent; in the time spent aboard ship following the comple- tion of training, skills may change markedly in either direction. Shipboard experience may result in significant improvement; or, if the shipboard experience was of a type which did not furnish opportunity for practice, the skill may definitely deteriorate. In the case of a radioman, for example, proficiency in receiving code may drop significantly even in the time intervening between completion of training and assignment to a billet, unless provision is made for refresher drill.22 The use of performance tests aboard ship is of course possible, and for some billets per- formance tests perhaps should be given rou- tinely and the results recorded on the man’s service record, with the date, for use in later classification. In wartime, however, elaborate routines of test administration aboard ship may not be feasible, and a more convenient procedure for evaluating performance is needed. The use of rating methods would provide the means for making such evaluations. The Need for Training of Raters. The com- mon errors in ratings (halo effect, error of leniency, etc.) are well enough known that a detailed discussion of them is not necessary here. It is probably sufficient to say that ratings of proficiency in shipboard jobs will be useful in advanced classification only to the extent that care is taken in planning the rating scales and training the raters in their use. Only through careful work can ratings of acceptable reliability and validity be obtained. The training of officers or petty officers who make the ratings is essential in order to obtain valid and reliable ratings. Training should in- clude discussions of errors such as halo effect, overrating, and prejudice, as well as an ade- quate description of the distribution of abilities and of the scale itself. That the results of using rating scales without opportunity to train raters or to review and follow up the raters’ evalua- tions is likely to be of limited value is shown by one recent study;10 a factor analysis study of the rating scale traits revealed that the rat- ings were characterized by a great deal of halo, and that almost any one of the traits could be used to represent the whole of the ratings. This was in spite of the fact that the traits rated were apparently quite heterogeneous, including such traits as adaptability, courage, sociability, and dependability. RESTRICTED SUGGESTIONS FOR IMPROVEMENT OF ADVANCED CLASSIFICATION 131 In order to avoid such results as were ob- tained in the study referred to above, one should not only make adequate provision for training raters, but also one should take considerable care in planning the rating scale itself. In gen- eral, the traits to be rated should be as objective and well defined as possible. Probably a graphic rating scale, on which certain points are defined with descriptive terms, is in general most sat- isfactory. Each trait should refer to a single type of activity. The effort should always be made to have each man rated by a number of his supervisors or officers, so that the final score can be based on the consensus of opinion and not on one person’s judgment. Work Readiness Tests Even with a fairly elaborate procedure for entering performance test scores and ratings in the service records of enlisted men, there are likely to be cases where data needed by a classification interviewer are lacking or out of date. He may be unable to make an assignment without resorting either to records of aptitude test scores and the like or to trusting state- ments made by the interviewee regarding his own training and experience. Tests similar to the oral trade tests of the U. S. Employment Service, but based on Navy jobs, would be par- ticularly valuable in such situations. The Applied Psychology Panel has made a beginning in the development of tests of this type. One Panel project has developed such a test (termed a work readiness test) for dis- tilling plant operators.13 This test consists of 15 items which may be administered orally or as a written test. Two forms were prepared which are equivalent with respect to difficulty, content, and discriminating power. The test has satisfactory reliability. It differentiates men with watchstanding experience on distilling plants from men with nonengineering rates and also from men with engineering rates but no distilling plant experience. Apparently neither chance nor experience other than dis- tilling plant operation is likely to produce spuriously high scores. Use of the test thus furnishes a method of evaluating the claims of those who profess to be experienced in dis- tilling plant operation. The test is objective and economical of time, and the results can easily be interpreted in terms of assignment to train- ing or to shipboard duty. The availability of a battery of work readi- ness tests covering a variety of billets would be of considerable value to a classification in- terviewer ; such tests could be administered in cases where more direct measures of pro- ficiency are not at hand. The procedures for developing work readiness tests are described in an Applied Psychology Panel publication.15 13.3.2 Extension of Advanced Classi- fication Procedures In order for advanced classification to be of maximum effectiveness, it should involve more than the making of recommendations, which may or may not be followed, only at relatively infrequent stages in the career of the enlisted man. There should be throughout the Navy sufficient appreciation of the aims and methods of classification at least for sympathetic coop- eration with advanced classification. It would be still more desirable to have officer personnel at all ships and stations who are qualified through interest and training to undertake classification procedures as an extra duty if not as a full-time assignment. In addition, all other officers, who do not serve as classification officers in a direct capacity, should nevertheless know enough about classification to use sensibly the data in a service record for assigning or advising enlisted men, and to realize the im- portance of keeping service records complete and up to date. One of the Applied Psychology Panel projects made a survey, in connection with a study on the validation of selection procedures, of the extent to which classification data were avail- able to ships’ officers on destroyers in shake- down.8 It was discovered that Division officers charged with the responsibility of assigning men in most cases were not aware of the ex- istence of the Q-card, and of those who did know of its existence few were aware that it contained information which could be used in making assignments. As long as this situation remains typical, it means that classification RESTRICTED 132 ORGANIZATION OF AN ADVANCED CLASSIFICATION PROGRAM data are of extremely limited value. Not all problems of assignment, training, transfer, and the like which occur aboard ship can be handled through recommendations made at a shore- based classification center. The study also revealed that Q-cards were frequently not available, that test records on Q-cards which were available were frequently missing or incomplete, and that Q-cards were often found for men not aboard. In other words, considerable laxity in connection with classifi- cation procedures was evident. Steps have been taken by the Enlisted Classi- fication Section which, if followed up effec- tively, should lead to considerable improvement in the attainment of the objectives of advanced classification. The provision for training one officer and one yeoman from each ship (see Section 13.2.8) is particularly important in this regard. It might be desirable in addition to have classification officers attached to Precommis- sioning Training Centers whose duty is to serve as temporary “technical aides” to the executive officer of each new ship. Such an officer could serve on each new ship from the time the crew was processed until the ship reported for shake- down. During this time he could work with the officer designated as classification officer for the ship, assisting in problems of classification as they arise and incidentally providing indoc- trination for the officers and yeomen. This plan has the advantage of providing instruction at a time when it is especially needed and con- sequently at a time when motivation to learn should be high. For long-term effectiveness, the indoctrina- tion of officers in general can best be achieved by the inclusion of definite courses in personnel procedures at the U. S. Naval Academy. This would eventually ensure some understanding of the problems of classification among high ranking officers in the Navy, as well as among engineering, deck, gunnery, and other types of officers throughout the Service. Only when the specialized training of classification officers is supplemented by general understanding and cooperation among other officers can the im- portant goals of personnel classification be achieved. Only then can classification become a continuing process through which the service record becomes more and more valid as a device for making assignments and enlisted men con- tinue to be given training and billet assignments which make them of maximum value to the Navy. As has already been mentioned, recommen- dations from classification centers are occa- sionally ignored, and billets may be assigned by the personnel department on the basis of the alphabetization of names, for example. It is not uncommon for assignments aboard ship to be made by assigning men to those battle stations which are nearest their bunks. In such situations a classification officer can do nothing, since billet assignments are actually a function of the personnel department. Similarly, classi- fication officers have no jurisdiction over the crew of a ship; once a man has been assigned to a ship, the commanding officer has jurisdic- tion over his future assignments. As a result, it sometimes happens that the poorest men tend to be assigned to advanced training, since the commanding officer is loath to give up a good man. One ship may have more than enough good first class radiomen while another ship has none. When a request for transfer is re- ceived, the commanding officer may circumvent the purpose of the transfer by getting rid of the poorest of his first class radiomen. In de- fense of this situation it is often said that the officer in charge of his ship depends upon his crew for the safety of his ship and its success in combat. While no one would gainsay this position, the long-term values to the Navy of equitable distribution of men among its fight- ing ships and of selecting the best qualified men for training should not be overlooked. Considerations such as these might lead to the suggestion that the responsibility for as- signments within a particular region (such as a Naval district or a fleet command) be placed in the hands of a well-qualified classification officer. He would have authority to see that classification recommendations are followed and that men are equitably distributed among the units of the fleet, and to make sure that all directives and policies regarding classification are followed. By making use of methods of evaluating proficiency, the classification officer RESTRICTED SUGGESTIONS FOR IMPROVEMENT OF ADVANCED CLASSIFICATION 133 can objectively control the equitable distribu- tion and proper assignment of the men under his jurisdiction. Through such a method, which would neither be too centralized nor give too much autonomy to separate ships and stations, the classification procedures might be made to work to the greatest advantage to the Navy as a whole. i'3'3’3 Procedures for Coding and Filing Personnel Data In the days before radio, radar, sonar, loran, fire control equipment, and automatic weapons, naval warfare was a much simpler matter than it is today. The use of Navy ratings at such a time was probably a fairly adequate device for classifying and selecting personnel. With the increased specialization that has come with a highly mechanized Navy, however, the value of rates for selection is highly questionable. A first class gunners mate, for example, may be entirely capable of operating and maintaining .50-caliber Browning machine guns but be rela- tively ignorant of larger automatic weapons such as a 40 mm or 5"/38 gun; he may have never served on a ship which mounted the larger weapons. Similarly an electrician’s mate may be quite capable of ringing out and re- pairing lighting circuits but be ignorant of selsyn units and interior communications. One possibility is to increase the number of ratings and to add designators to ratings which indicate fields of specialization. This method has been used to a certain extent. New Navy rates, such as sonarman and radarman, have been added, and the designators (M) and (R) have been added to fire controlman (FC) to indicate whether the man is qualified for main- tenance or for operation of fire control equip- ment. This is at best only a partial solution, however, unless the system is extended consid- erably. For example, in assigning a fire con- trolman (M) to a particular billet, it might be necessary to know his qualifications for maintaining Mark 37, Mark 51, and Mark 52 gun directors, rangefinder optical equipment, torpedo directors, fuze setter and sightsetter equipment, and other types of computers, stable elements, and the like. The problem is just as complex for many other rates. A machinist’s mate may have specialized in high or low pressure steam turbines, reciprocating steam engines, diesel or gasoline engines, or turbo- electric equipment. A radarman may be quali- fied on only one of a number of different types of radar sets. In other words, it would seem that what is needed is a method of coding Navy occupational skills which can be used efficiently in classifica- tion procedures. A five- or six-digit code would probably suffice, in which various digits or pairs of digits could be used to represent rate, pay grade, field or fields of specialization, and degree of proficiency in the field of specializa- tion. Classification would be facilitated by ap- propriate use of such codes both at shore-based stations and on ships. At shore stations the code could be punched into electric accounting punched cards, and sorting of the cards repre- senting the men in a pool could be quickly made. This would rapidly make accessible the data needed for making billet assignments and would require much less individual study of Q-cards. For shipboard use,10 the code would permit the use of such a device as the McBee Keysort sys- tem.0 The Classification and Selection Section of the Bureau of Navy Personnel23’24 has de- veloped a Manual of Navy Enlisted Job Classi- fications which provides a Navy occupational skills code. This code went into effect July 1, 1946. RESTRICTED Chapter 14 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS Norman Frederiksen a Summary This chapter consists of an outline of the procedures recommended for constructing aptitude tests for the selection of military per- sonnel. 141 INTRODUCTION The procedures which the Applied Psychology Panel used and recommends for use in prepar- ing tests for the classification of military per- sonnel are described in this chapter. They apply particularly to the development of aptitude tests of the paper-and-pencil variety which are power tests rather than speed tests. The pro- cedures are described here in detail because they are not generally available. They are de- scribed with refinements which could not always be used because of the speed demanded by war- time pressure. The best tests constructed under Applied Psychology Panel direction, however, were those in which the recommended proce- dures were followed most closely; when short- cuts were used, the tests suffered by comparison. In peacetime test construction, thoroughness in the application of the methods described is strongly recommended. It is impossible to construct valid tests merely by sitting at a desk and writing test items. Be- fore any work on test preparation is done, it is necessary to find out what the proposed test is to be used for—what it is supposed to predict. It is necessary to study the to-be-predicted ac- tivity in order to determine the critical skills and abilities required for success. It is further necessary to acquire at least a general notion of the characteristics of the group to which the test is to be given, and to become acquainted with the conditions and limitations of the clas- sification program of which the test is to be- come a part. Only after these things have been done can one start writing test items. The items them- selves must then be tried out and an item analysis made. It is necessary to select items for the final test in terms of their difficulty and validity, to determine time limits, to make cer- tain that instructions are clear, and to arrange the items in proper order. All these things are necessary before one can say with confidence, “Here is a good test.” 142 PRELIMINARY DETERMINATION OF WHAT THE TEST IS TO ACCOMPLISH 14-21 The Performance to Be Predicted The first task is to decide specifically what performance is to be predicted. This perform- ance may be something as specific as using a stereoscopic heightfinder, or something as gen- eral as the intellectual ability to pass any Serv- ice school course. Determining What Is to Be Predicted The most general problem likely to arise is whether a classification test battery should be aimed at predicting Service school success (the immediate goal of trainees) or eventual success in carrying out a military job. This problem arose in connection with the preparation of the Navy’s Basic Classification Test Battery. The decision made then was to attempt primarily to predict Service school success, on the grounds that it was especially important to prevent waste of time and equipment in trying to train men who were poorly qualified with respect to this first goal. At the time the decision was made, the choice may have made some differ- ence in the types of tests which were developed; but with improved, more realistic training in Service schools, such that the activities in Serv- ice schools resemble more closely the duties 14.1 a This chapter is based on the work of NDRC Project N-106 and the procedures followed by recognized test experts. 134 RESTRICTED PRELIMINARY DETERMINATION OF WHAT THE TEST IS TO ACCOMPLISH 135 aboard ship, it would make less difference which criterion one aimed to predict. Determining the Test Content After deciding what is to be predicted, it is necessary to select the kind of material to in- clude in the test. The second step is to make a thorough job analysis of the performance to be predicted. (See Applied Psychology Panel, Volume 2, Chapter 14.) In some cases the re- sults of a job analysis will already be available. These analyses should be studied and used, but acquiring firsthand information about the job should not be neglected. If a test is to be made to predict success in a signalman school, for exam- ple, the test specialist should visit such schools, study the curricula, analyze the causes for fail- ure, and if possible actually take the course him- self to find out at first hand what the problems are like. Only by such firsthand experiences can he really get the feel of the situation and the background which best equips him to form hunches as to the types of items most likely to be successful. Evaluation of Training. At this point it may be mentioned that the test specialist has an opportunity and a responsibility to make evalu- ations which have implications beyond the field of classification. He may find that training is not realistic and that tests constructed to pre- dict school success will actually select men poor in performance of their military duty. For ex- ample, one school curriculum led to the selection of bookish men, while a quite different type was most successful on the job. (See Chapter 17 of Volume 2, Applied Psychology Panel.) Obviously in such a case it would be extremely shortsighted to go ahead and prepare a test which would predict only school success. Instead steps should be taken to make the training more realistic, so that both training and selection are better oriented toward the ultimate goal of maximum quality of performance in combat. The Availability of a Criterion Before beginning the work of test construc- tion, plans should be made for validating the test. Validation requires that a reliable and meaningful criterion of success on the job be available. In many cases it will be found that a suitable criterion is lacking or that the avail- able criterion (such as passing or failing a Service school course) is itself probably not a valid indication of success. If a suitable criterion is lacking, it will eventually be necessary to develop one. If the development of the criterion measure is not begun until after the predictive test has been completed, the validation of the test will be considerably delayed. Serious consideration should therefore be given to the possibility of beginning work on the development of a cri- terion measure at the same time that the con- struction of the test is started. 14.2.2 Analysis of the Classification Situation The Testing Situation In addition to studying the performance in which success is to be predicted, it is also de- sirable to investigate the situation in which the proposed tests are to be used. Factors such as time available for test administration and test scoring, size of groups to be tested, use of interviews or other classification procedures, and physical facilities for conducting a testing program have a significance with respect to the nature of the tests to be developed. Considerations such as those suggested above should not be allowed to have too much influ- ence in determining policies with regard to test construction. During peacetime develop- mental work, it should be possible to relegate such factors to a secondary role and to shape the situation in which the tests are to be used to fit whatever tests are found to be useful. On the other hand, one should guard against set- ting up a testing organization which, however suitable it may be for peacetime needs, is too elaborate and time-consuming to permit classi- fication of recruits at a rate which might become necessary during mobilization for war. Characteristics of the Group to Be Tested Advance information concerning the charac- teristics of the group to whom the test is to be given will be of value in designing tests and RESTRICTED 136 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS will help in gauging the proper difficulty of test items. Knowing the central tendency and vari- ability of the group with respect to age, educa- tion, vocational training, and occupational ex- perience should be of value in this connection. Score distributions and item analyses of tests previously taken by the group will be particu- larly useful if they are available. 14.3 PREPARATION OF THE FIRST EXPERIMENTAL TESTS The first version of a test should be thought of as purely experimental; it should be designed to give as much information as possible to aid in future revisions of the test. This section in- cludes specific suggestions with regard to planning the first experimental version of a test for maximum usefulness in obtaining data for subsequent revision, 14-31 Writing the Test Items On the basis of the preliminary survey of the performance to be predicted, the psychologist can prepare the items included in the first ex- perimental test. The first version is in the na- ture of an experiment, and all reasonable hunches should be followed up. This means that not one, but several tests or subtests should be constructed, with the idea of following up the best leads suggested by the preliminary re- search findings on comparative validity, relia- bility, and internal consistency. The findings reported in the literature by previous investi- gators should be studied for ideas with respect to suitable test items. Now for some more specific suggestions with respect to the preparation of the first experi- mental tests: 1. Make three to five times as many items as will be used in one test. A sufficient number of items should then prove satisfactory to permit the selection of matched items for at least two parallel forms of the test. For some kinds of items (such as arithmetical computation) where comparatively few of the items will fail to meet the standards for selection, fewer items need to be prepared than for types of tests where successful items are more difficult to write, such as verbal analogies and mechanical comprehension. The number of items that can be included in one test is limited, of course, by such factors as fatigue and boredom of the testees. 2. Edit the items carefully. A lower mortality rate will result. Too often it is found that an item which must be rejected because of item analysis results contains a defect which could have been discovered in advance. The editing should include not only careful study by tech- nical experts (such as the study of mechanical comprehension items by engineers) but also careful inspection from the point of view of rhetoric and ambiguity. 3. Strive for a fairly wide range of difficulty in order that items can be selected which have a suitable range of item difficulty. Information on the general ability level of the group to be tested will help in this connection—especially information on tests of similar nature which have been administered to that group. Experi- ence has shown that it is harder to construct valid difficult items than valid easy items;22 therefore a greater excess of difficult items than of easy items should be prepared, in order to have available a sufficient number of valid diffi- cult items. 4. Choose the type of item—multiple-choice, matching, true-false, completion, or other type —which appears to be most advantageous in the particular situation. In making the choice, consider such factors as time and space per item, ease and objectivity of scoring, freedom from chance success, and ease of giving direc- tions, as well as the appropriateness of the type of item for the kind of subject matter being dealt with. All things considered, the multiple-choice type is in general most satisfactory. Tests com- posed of multiple-choice items are easy to ad- minister, they are adaptable to machine scoring or are easily scored by hand, and with four or more choices are sufficiently free from the fac- tor of chance. A large number of such items can ordinarily be administered in a short time. Most important, the multiple-choice item is adaptable to a wide variety of purposes, from RESTRICTED PREPARATION OF THE FIRST EXPERIMENTAL TESTS 137 measuring simple recognition to complex think- ing. However, for certain kinds of subject mat- ter other types of items may be superior. That type of item should be chosen which is best adapted to the type of measurement being at- tempted. 5. For each item, prepare two or three more distractors (wrong answers) than are planned for use in the final form of the test. It will then become possible to select those distractors which function best. 6. Base the selection of distractors on objec- tive empirical evidence when feasible. In order to do this, the psychologist should prepare a free-answer form of the test (i.e., a completion or short-answer test in which each testee writes his answer to each item). The free-answer test should then be administered to a group similar to the population eventually to be measured. The frequency with which various answers are given can be tabulated and the most frequent wrong answers selected as potential distractors for the multiple-choice form of the test. By use of a more elaborate item analysis of the free- answer test, it is possible to select items and distractors on the basis of their discriminative value. It is further possible to estimate the difficulty of each item on the basis of the fre- quency of correct answers. This method is obviously not applicable to all types of tests; it could not, for example, be used to good advantage on a test where the choices are represented by pictures. The method is applicable when there is the greatest simi- larity between the operations performed in a free-answer and a multiple-choice form of the test. Arithmetic or mathematics tests are espe- cially adapted to this technique of obtaining distractors.21 7. Try to present the test material as real- istically as possible; for example, the material in arithmetical reasoning items should be con- sistent with the experience of the testees to as great an extent as possible. Use pictures when- ever they are appropriate. When testing a group with a wide range of ability, it is difficult to avoid high intercorrelations among tests. By avoiding verbal items except where they are especially appropriate, the intercorrelations may be reduced. Planning the Format and Directions A group test cannot be successful unless the directions for taking it can be followed. Many testees will have had little previous experience in taking examinations and will have a rather low level of ability in following either oral or written directions. It is important that a man’s test score represent his ability with respect to what is being measured by the test and not ability to follow complicated directions on cod- ing correct answers or using separate answer sheets. The format of the test and the oral and written directions should therefore be as sim- ple as possible, and format and directions should be studied during the use of the experimental versions with their later improvement in mind. Format of the Test All details should be carefully studied from the point of view of ease of understanding and following the test procedures. The choices for an item should be arranged in a column instead of being strung out along the page, even though more space is required. This arrangement makes clearer what identifying letter or num- ber goes with each choice. If the items are numbered, the choices should be lettered, so that in recording answers there is a minimum of confusion and error; (item) U, (choice) C is easier to bear in mind while recording on a separate answer sheet than U, 3, which might easily be recorded as 3, U- Use of Separate Answer Sheet Probably it will be decided that the test should be scorable by means of a test scoring machine; if so, the test blanks for the prelim- inary forms of the test should have answer sheets which are good facsimiles of, if not actual, machine scorable answer sheets. The machine scoring feature does not necessarily involve the use of a separate answer sheet; in some cases, it is feasible to print the test itself on the answer sheet.21 This would eliminate the necessity of coding the choices and making answers according to this code, which is un- doubtedly a source of some error. If the entire test can be printed on the answer sheet, this RESTRICTED 138 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS method definitely should be employed. If sev- eral pages are required, however, it becomes necessary to add scores from different pages, which introduces the likelihood of large errors; consequently the separate answer sheet would be preferable in such a case. Writing the Directions The writing of clear yet concise directions requires considerable care and insight into the thinking of the none-too-bright recruit who has perhaps never before taken a standardized test. The directions should be written in such a way that the test procedures can be correctly fol- lowed by all recruits—the dull as well as the bright. Test scores should not reflect differences in ability to understand directions. Perhaps the best approach in writing test directions is to describe, one at a time, the specific procedures which must be followed in the order in which they would most naturally occur. The recruit can be carried through the procedures most effectively by requiring him to participate actively through answering and recording answers to several practice questions. Sentences should be short, and vocabulary should be at a level which is low enough for all but the most illiterate to understand. In most cases, and especially for tests which ne- cessitate rather complicated procedures (such as Surface Development), it will be of consid- erable help to pretest the directions, using a sample group which includes many low-ability recruits. The directions should be very explicit in re- quiring the testee to answer every question. If this is not done, it will be necessary to cor- rect the scores for differences due to chance success. Even then, final scores will depend in part upon the men’s willingness or unwilling- ness to guess. The men are more likely to follow a common policy if guessing is encouraged than if it is discouraged. The word guess may be undesirable, but the same result can be achieved by saying, “If you don’t know the answer, mark the one you think is most likely to be right.” Making the Experimental Tests of Maximum Value A number of precautions should be observed in planning the first experimental tests to make the item analysis data of maximum value. In the first place, the format of the test and an- swer sheet should be such that part scores are readily obtainable, so that items of different types can be studied separately. Items of dif- ferent types can later be thrown together in an omnibus form, for example, if desired; but it is important to be able to study separately the efficiency of items of various types. Further- more, it is important to attempt to plan the test in such a way that data on the various items are equally trustworthy, i.e., affected as little as possible by such factors as fatigue, monotony, rate of work, or attitude. One way to minimize the effects of such factors is to publish the experimental tests in two or more forms which differ in the order of the items. Thus items which are near the end of the test in one case will be near the beginning in an- other. Ideally, the time limit should be so liberal that at least 90 per cent of the recruits are allowed to finish; otherwise the data available for the terminal items are likely to be based on a select group. Unless time limits can be deter- mined by pretest experiments, it may be desira- ble to leave the time limit undecided until the test is actually given. Then the test can be terminated when the desired proportion has finished. The number still working at any mo- ment can be determined by requiring the men to indicate in some way, such as by putting the test paper on the floor, when they have finished. 144 ADMINISTRATION AND ANALYSIS OF THE FIRST EXPERIMENTAL TESTS 14.4.1 Administration of the Test After the test booklets and answer sheets for the experimental test have been prepared, the test is ready for its trial administration. The sample to which it is given should resemble as much as possible the population with which the test is ultimately to be used; otherwise esti- mates of item difficulty (in particular) may be misleading.18 In order that the various sta- tistics used to evaluate test items may have satisfactory reliability, the sample should be made up of a minimum of 500 men. If different RESTRICTED ADMINISTRATION AND ANALYSIS OF THE FIRST EXPERIMENTAL TESTS 139 forms of the experimental test (containing the items in different order) are used, they should be given to subgroups of equal size selected at random from the sample. The tests should be well proctored, and the proctors should be in- structed to notice particularly behavior of the testees which indicates how successful the test directions have been. Questions by the recruits would be such an indication of the adequacy of the directions. A record should be kept of any such evidence which is forthcoming. The test should not be terminated, if possible, until at least 90 per cent of the men have completed the test. 14.4.2 Item Analysis of the Test The test is now ready for tabulation of re- sponses and item analysis.19 The analysis of the test should furnish for each item some measure of its difficulty and some measure of the validity of the item (i.e., its relation to some criterion, usually the score on the test or subtest of which the item is a part). In addition, the item analy- sis should show how many individuals chose each alternative answer to each item and the average score (on the criterion) of the group choosing each alternative. Finally, the relia- bility of the test should be determined by corre- lating random (usually odd-even) halves of the test and correcting for length by means of the Spearman-Brown prophecy formula. Once these computations have been completed, the first revision of the test can be made. The Analysis Procedure The amount of item analysis information available about each item and about each alter- native within an item determines the adequacy of the revisions that can be made. A very de- tailed item analysis procedure19 was, there- fore, followed in constructing the Navy tests described in Chapters 2 and 3.b Figure 1 shows a sample item analysis sheet. It gives the information which was regularly used in making final choice of which items to include and which to reject. The types of infor- mation made available and their interpreta- tions are: Item. The item itself is typed on the sheet. For ease in sorting, its number is also printed in the box at the upper left. Distribution of Responses. The number of men selecting each alternative and the number omitting the item entirely are shown in the column headed “N.” Three men omitted item 96. Two hundred and eighty-eight answered it correctly by choosing the first alternative. The heavy lines above and below row 1 indicate that that was the correct answer. The other 209 men failed the item. Rows 2, 3, and 4 supply information on how many chose each wrong alternative. Number of Cases. The Base N is given at the top. The number who actually attempted an item (Nf) appears in the bottom row of the table. Anyone who has answered a particular item, and anyone who has answered a subse- quent item, is considered to have tried that item. The sample item shown in Figure 1 was placed early enough in the test so that Nt equaled Base N. In items placed near the end of a test, Nt is frequently less than Base N. Nt rather than Base N is used in computing the statistics described in the following para- graphs. Item Difficulty. Two measures of item diffi- culty are used. The simpler is symbolized by p. It consists of the percentage of Nt which an- swered the item correctly. In this case p = .58. The second difficulty measure is symbolized by the Greek letter delta (A). A is expressed in terms of transformed criterion scores. The essential features of these transformed scores are that they correlate 1.00 with the original scores, that the mean of the total sample on the transformed scores is uniformly 13.00, and that the standard deviation of the total sample is uniformly 4.0. A is defined as that trans- formed criterion score above which the percent- age of cases equals p. The more difficult the item, the higher the value of A. In this sample A = 12.2 and p = .58. For an item with a A value of 13.0, p would equal .50. Correlation with Criterion Score. The bi- serial correlation between the item and the criterion is computed. In this case it equaled .60. b Reference 19, Characteristics and Uses of Item- Analysis Data, has been declassified and will be pub- lished by the Stanford University Press in the series of Applied Psychology Monographs. RESTRICTED 140 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS COLLEGE ENTRANCE EXAMINATION BOARD Research and Statistical Laboratory Princeton, New Jersey ITEM ANALYSIS Card Number 48 TEST ME CHKN FORM j BASE N 5 0 0 Date Tabulated 6 44 Operator Number 9 Response Code n 2x Mean Sx2 n+ /Sx2 P = a =,/ M3 "t K n '■(^XO 96. In a gasoline englne,_ the gas mixture should explode In — (a) the cylinder (b) the intake manifold. (o) the exhaust manifold. (d) the car- buretor. NOTES: — 1. x1, based on p, is the distance from the mean along the baseline of the normal curve, in terms of unit standard deviation. If p is less than .5, x' is positive; if p is more than .5, x' is negative. 2. Compute means only for responses made by ten or more candidates. 3. Compute M+ and Mt to three decimal places; means of all other responses, to one. 4. Record r to two places;^to one. 2xz 2xz 5. Carry all other computations to three decimal places; i.e., , n n Figure 1. Sample item analysis sheet. RESTRICTED ADMINISTRATION AND ANALYSIS OF THE FIRST EXPERIMENTAL TESTS 141 The criterion employed is the score on the sub- test of which the item is a part; if the test is not divided into subtests, then the score on the total test is employed. Information on Alternatives within Each Item. The mean transformed score of the men selecting each alternative is given in the col- umn headed “mean.” The mean score of the 288 men who selected the correct alternative was 14.635, substantially higher than the mean of the entire group (13.0). The mean transformed scores of the men who chose the wrong alterna- tives were all well below 13.0. Alternatives c and d (mean scores 10.2 and 10.4) appealed to a poorer group than alternative b (mean score 11.9). Alternative Criteria for Item Analysis If it is possible to administer the trial test to a group which satisfies the criteria for a representative sample and at the same time can be used as a validation group, not only can the test be validated more quickly, but also a supplementary item analysis using the external criterion can be employed; this permits the selection of items on a dual criterion: internal consistency and correlation with the external criterion (see Chapter 15). Still another type of analysis would be of value in situations where, as in developing a battery of tests, it is desirable to reduce the in- tercorrelations of the tests as much as possible. This cross item analysis consists of correlating the items of each test with scores on all the other tests in the battery, thus permitting the selection of items on the basis of high correla- tions with their own test and low correlations with other tests. Such a procedure would be particularly useful for recruit tests, where the range of talent is great, and for tests where low correlations are desirable but difficult to obtain, as in tests of verbal facility and arith- metical ability. Selection of Items for Alternative Forms of a Test It is always desirable to have two or more alternative forms of any test which is to be used widely and repeatedly. The construction of alternative test forms always poses a num- ber of statistical problems. The distributions of scores on the two forms should be similar. The difficulties of the items should be similar. Item- test correlations, reliabilities, and other statis- tical characteristics should be matched for the two forms of the test. The satisfaction of these requirements requires careful planning and considerable work. The procedures to follow in developing equivalent forms of a test are in general those described in this chapter. References 17 and 18 describe in detail the application of these pro- cedures to the special problems of developing equivalent test forms. Uses of Item Analysis Data The principle uses of item analysis data may be briefly summarized as follows :19 1. Item analysis supplies detailed, objective, quantitative information for each item. This in- formation cannot be obtained by “expert judg- ment” nor by any manipulation of the relia- bility coefficient. 2. The objective, quantitative information from item analysis is well suited to help settle arguments or objections concerning specific items; and it provides a convenient, practical basis for selecting items for subsequent forms of a test. 3. Item analysis data provide information which is useful in revising and improving test items. 4. The distribution of item difficulty can be improved with respect to symmetry, continuity, and average level, on the basis of the evidence provided by item analysis concerning the diffi- culty of each item. 5. The reliability of the test may frequently be improved by the judicious selection of items on the basis of item analysis data. 6. The independence of the test from other tests in the battery may be improved by the application of a cross-item analysis technique. 7. The external validity of the test can fre- quently be improved, if the item analysis in- cludes the correlation between each item and a valid external criterion. 8. Item analysis data stimulate hypotheses and insights useful in construction of tests and interpretation of test results. RESTRICTED 142 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS 145 REVISING THE TEST The revision of the test is a matter of select- ing items on the basis of the item analysis statistics and modifying the format and direc- tions. 14 51 Selecting Test Items Determining the Number of Items Needed On the basis of the reliability coefficient which was obtained for the trial test, it is possible to determine about how many items should be in- cluded in the test. Ordinarily a reliability of about .90 is desirable; the number of items necessary to secure this reliability can easily be estimated by means of the Spearman-Brown formula. Various other considerations may of course lead one to revise that number in one direction or another. It might, for example, be desirable to decrease the number of items some- what in order to avoid excessive time for test administration or to avoid using more than one side of an answer sheet. It should be remem- bered that the estimate of the reliability of the test will probably be an underestimate, since the selection of test items should ordinarily re- sult in a test which is improved with respect to internal consistency and reliability. Criteria for Item Selection Items are then selected on the basis of a multiple criterion which includes the difficulty of the items and their relationship to the total- test score. The most desirable items are those which have the highest correlations with the total-test score and whose difficulty values are scattered near a level halfway between a chance level and the level where all testees answer the item correctly23 (e.g., for five-choice items, those which are passed by about 60 per cent of the group). The most desirable scatter about this level of difficulty is not entirely clear as yet; but from practical considerations it would seem desirable to include a few relatively easy items as “icebreakers” at the beginning of the test and to avoid an excessive number of items which are so difficult as to have an ad- verse effect upon the attitude of the testees. For a five-choice test, the items should probably range in difficulty from those answered cor- rectly by about 85 per cent to those answered correctly by about 35 per cent, with the majority of the items at about the 60 per cent level. Assuming a fairly narrow distribution of item difficulties, the optimal values for item- total test correlation is a question not fully an- swered as yet. Presumably no serious errors will be made if one selects the items with the highest correlation with the total test, pro- vided a sufficient number of items are avail- able whose biserial correlations are above, say, .25. If more than enough items are available which are fairly homogeneous, as indicated by uniformly high biserial correlations, it is pos- sible to select on a judgmental basis, choosing those items which best represent the hunch being tested. The lower the item-total test correlations tend to be, the less value they have for item selection. If they are all extremely low, they indicate merely that the items are heterogene- ous. In such a case items can be selected on the basis of subjective judgment as to similarity. A better technique would be to subject the items to a factor analysis and select items on the basis of their loading with a particular fac- tor. The factor analysis might be based on a subjective classification of the items into groups which appear to be alike; the scores on the groups, treated like separate test scores, could then be correlated and factor-analyzed. Selection of Distractors It was also recommended previously that the items of the trial form should contain more distractors than are needed for the final form of the test, in order that the best distractors can be chosen. The recommended item analysis procedure would supply the necessary infor- mation, viz., (1) the number of people who choose each distractor and (2) the mean score on the test for each group. The most efficient distractor is one which attracts a fairly large group of low-ability men; therefore those dis- tractors should be eliminated which relatively few choose, and those which are chosen by high-ability men more frequently than by low- ability men. The relative importance of these two factors is difficult to evaluate in some instances; for example, it might be difficult to RESTRICTED REVISING THE TEST 143 decide whether to eliminate choice a or b in the following example: through their incorporation in another experi- mental form of the test. The Use of External Criteria in Item Selection The above description of the procedures for selecting test items will perhaps be adequate for the situation in which the test is analyzed using only total-test score as the criterion. With an external criterion, a somewhat more elaborate procedure of item selection must be used.15 This method is described in Chapter 15. Use of an external criterion is most useful when there is only one specific criterion to be predicted; in cases where success in any school or any of a limited number of schools is to be predicted, the method would be less practicable. Where the items of a test have been corre- lated with other aptitude tests as well as with the test of which the item is a part, it is pos- sible to minimize the intercorrelations among the tests of a particular battery, by selecting items which have high correlations with their own test and low correlations with the other tests. 14 5 2 Revising the Directions and Format Revising Directions If one has been careful to observe the per- formance of the recruits with respect to their understanding of the test directions, the modi- fications which are required should be appar- ent. In scoring the tests it may be noticed where common errors are made, such as recording answers in the wrong column of the answer sheet or confusing the spaces for practice ques- tions with spaces for the test proper. If evi- dence of such errors is found, appropriate changes in format or directions should of course be made. The revisions might take the form of modification of written directions, boxing-in the practice questions and answer spaces, or changing the method of coding an- swers. Determining Time Limit It may be possible at this point to determine precisely the time limit for the test, on the basis of observations of rate of work on the Number Mean score of choosing those choosing Alternative each alternative each alternative a 37 14.1 b 14 13.0 c 31 10.6 d 24 10.4 e 62 12.3 Correct answer; f 116 17.3 Alternative a is chosen by a considerably larger group than b, but those who choose a make scores which are on the average considerably higher than the scores of those who choose b. (The means are in terms of standard scores based on a distribution with a mean of 13 and a standard deviation of 4.) No rule can be presented to govern a situation of this sort; further research, dealing with different types of test material, is needed. Revising Items In cases where the experimental test fails to yield a sufficiently large supply of items or when additional forms of the test must be pre- pared, the rejected items should not be over- looked as a source of supply.19 Since a con- siderable amount of statistical information concerning the item is available, the chances of finally achieving a satisfactory item through revision are often superior to the likelihood of writing a successful new item. The data may show, for example, that some distractors are successful—they draw a goodly number of men whose scores are on the aver- age well below the average for those who got the item right. Such distractors should be re- tained. The failure of the item may be due to a distractor which draws a high-ability group; if so, it may be sufficient to rewrite this dis- tractor. In other cases it may be necessary to rewrite distractors to make them somewhat more attractive, so that they will become func- tional parts of the item. It is also possible, of course, to revise the part of the item preceding the choices; this would mean that the statistical information would be invalidated. After revi- sion, the items are ready for a second trial RESTRICTED 144 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS experimental test. More likely, however, it will be necessary to establish the time limit on the basis of further experimental administrations of the revised test to representative samples of the population. In the case of a power test, the time limit should be such that additional time would not affect the test scores of students to any great extent. Arrangement of Items It was recommended above that the number of items in the trial test should exceed the num- ber required for one test by a considerable amount, in order that parallel forms of the test could be prepared with a minimum of effort. The two forms of the test should contain items which are matched with respect to con- tent, difficulty, and correlation with total score, and should resemble each other as closely as possible with respect to means of item difficulty values. If sufficient items which satisfy the above criteria are not available, of course only one form of the test can be obtained, and more work will be required in order to try out addi- tional items. The items which have been selected for each form should be arranged in order of difficulty as determined by the item analysis. This ar- rangement will permit a somewhat shorter time allowance than would otherwise be necessary, since there is ordinarily a sufficiently high cor- relation between level of ability and rate of work so that the men who have no opportunity to attempt the last items would be unable to answer them correctly anyhow. Having the items in order of difficulty has the advantages of permitting the men to build up confidence at the beginning and enabling them to attempt the majority of items that they have the ability to answer. With items in a different order, the men will waste time trying to answer items which are too difficult for them while leaving easy items unattempted. Preparing a Test Manual The revision of the test is not ready for gen- eral use until a manual of directions for ad- ministering it has been prepared. This manual should be a very explicit step-by-step descrip- tion of the preparations and procedures in- volved in giving the test. The importance of uniform procedures has been shown in various studies8-13> 22 which reveal differences in test results from station to station, differences apparently due to variations in procedure. The manual should include such topics as suitable conditions for test administration, specific duties of proctors, distributing test booklets, methods of keeping an adequate check on the number of booklets, and seating arrangements, as well as a specific account of the steps to fol- low in administering the test itself. All words that are intended to be read aloud by the ex- aminer should be printed in boldface type or otherwise set apart typographically in order that they can be found readily at the proper moment in giving the test. The test manual should also contain detailed and specific direc- tions for scoring and for converting raw scores to whatever units are to be used in reporting test results. 146 VALIDATING THE TEST If the test was not validated at the time of its first experimental administration, it is now ready for validation against the criterion which it was designed to predict. The results of the validation study determine which of the vari- ous experimental tests that have been tried out will be adopted and put into general use. They also may suggest new hunches with respect to types of tests which may be prepared and tried out. This experimental administration also fur- nishes opportunity for a statistical evaluation of the test in its revised form. The reliability of the test may be checked, the items and their distractors may be re-evaluated, and the time limits and test directions may be re-examined. A new item analysis would be particularly im- portant for evaluating items from which extra distractors have been eliminated, since there is always the possibility that the distractors which remain may not function in exactly the same way when a change has been made in their context. After any further revisions have been made, the test is ready to put into gen- eral use. RESTRICTED ESTABLISHING NORMS 145 14 7 ESTABLISHING NORMS i4.7.i Adequacy of the Sample In establishing norms for a test, the most important consideration, so far as early plan- ning is concerned, is the adequacy of the sample. The sample should be large, in order to avoid sampling fluctuations. It is difficult to state a minimum number, because this would depend to some extent on the number available; but upwards of 2,000 cases would be satisfac- tory in most instances. More important than size is the representa- tiveness of the sample. Perhaps the best pro- cedure would be to take all cases to which the test would routinely be administered during a particular interval of time (say one month) during which no selective factors are known to be operating.7-11 14-7 2 How to Report Scores The next problem to be considered is the type of measure which should be used in re- porting scores. A desirable measure should possess the following characteristics. 1. It should be easy for the layman to in- terpret. 2. It should have a standard interpretation, regardless of what particular test is involved. 3. It should not be based on units which are obviously unequal (i.e., a difference of five points near the end of the distribution should mean the same as a difference of five points near the middle of the distribution). 4. It should not be unduly influenced by minor irregularities in the shape of the distri- bution and should be relatively stable over a period of time. 5. It should involve the use of not more than two digits, for economical use of space on the punch card. 6. It should lend itself to use with available statistical procedures. Raw scores obviously are unsatisfactory for reporting test scores, since they have no stand- ard interpretation and cannot be directly com- pared from test to test. Percentile rank is sometimes used to indi- cate the relative position of particular scores in the distribution. Percentile rank has the advantage of easy interpretation by the lay- man, but its disadvantages are too numerous to justify its use: it has unequal units, it is influenced by minor irregularities in the shape of the distribution, and it does not possess arithmetical characteristics which permit its use in statistical procedures. Some form of standard score most nearly meets the specifications outlined above. Its most important disadvantage is its unfamiliar- ity to the layman, making interpretation diffi- cult. There is nothing intrinsically difficult, however, in the concept of standard score, and most officers, with a slight amount of indoctri- nation, can readily learn to use these scores. Standard scores satisfy the requirement of equal units well enough for all practical pur- poses, provided the score distributions do not depart too markedly from the form of the normal curve, and they do lend themselves to statistical manipulation. By using standard score units based on a mean of 50 and a stand- ard deviation of 10, all standard scores can be expressed in two-digit numbers. The Navy standard scores which are now in use are of exactly this type. Army standard scores, which have a mean of 100 and a standard deviation of 20, require three-digit numbers for reporting half the scores but possess all the other ad- vantages listed. The experience of World War II demonstrated that standard scores can be used by nontechnical officers. Preparation of Conversion Tables Conversion tables should be prepared, after the test has been scored, for the purpose of affording a convenient method of transmitting raw scores to standard scores. The conversion table should consist of a listing of all obtainable raw scores, each of which is paired with its corresponding standard score. The conversion table should be included with the manual of directions for administering and scoring the test. A Supplementary Method of Reporting Scores For the benefit of officers who are untrained in statistical procedures yet who will need to RESTRICTED 146 CONSTRUCTION AND STANDARDIZATION OF GROUP TESTS make use of test scores to assist in classification problems, it may be desirable to furnish a supplementary method of reporting scores which is simpler and coarser than the standard scores. The method might consist of reducing the scale to about five categories which are de- fined in terms of percentage of the population at various regions along the continuum. The Enlisted Classification Section of the Bureau of Naval Personnel has included in its recom- mendations for a postwar classification pro- gram the suggestion that such a scale consist of five categories, as follows. 1— top 7 per cent 2— next 24 per cent 3— middle 38 per cent 4— next 24 per cent 5— bottom 7 per cent Such a scheme used in conjunction with stand- ard scores should come very near to satisfying the criteria listed above. It is important that standard scores be included in each man’s rec- ord, however, because of their potential value for research studies. A five-point scale would be very inadequate for correlational studies, even if all categories were represented in a par- ticular sample; and it is very likely that, because of selection, only two or three of the categories would be represented in a particular group. i4.8 CONTINUATION OF TEST DEVELOPMENT The work of test development cannot be con- sidered completed at this point, except in a very restricted sense. A test constructed according to the procedures outlined in this chapter will be a good test at the time of construction; but there is no guarantee that it will continue to be a good test as time passes. Service require- ments may change as new equipment and tac- tics are developed, and the content of training programs may be modified accordingly. As a war progresses, the quality of the available manpower may also change. For these and vari- ous other reasons, a continuing program of re- search is necessary. Tests should be re-evaluated periodically, particularly from the point of view of validity. Norms should be checked occasionally in order to be aware of any shifts which may occur in the character of the population being tested. The search for improved testing methods and more valid types of test items should be a con- tinuing process. Only through such a continu- ing program of research can the tests be kept up to their maximum value in classification of military personnel where the importance of tests in maintaining efficiency is recognized. RESTRICTED Chapter 15 CONTRIBUTIONS TO TEST THEORY AND TEST ADMINISTRATION By Dael Wolflea Summary An investigation of the accuracy with which u aptitude tests were being scored in 1942 in the Navy revealed frequent and large errors. Methods of controlling and checking scoring procedures were devised. Subsequent investiga- tions showed that errors were greatly reduced in frequency and in size. Several innovations in test construction prac- tice were tried out. They included: 1. A simple method of measuring the pre- mium which a test places on speed. 2. A method of estimating the validity of a test for a school population without waiting for training and performance measures to accu- mulate after giving the test to the population. Instead, the test is given to a graduating class and their test scores are correlated with final grades. 3. A method for increasing the validity of a test by selecting items in terms of their corre- lation with an external criterion such as suc- cess at the school, instead of correlation with the total-test score. 151 INTRODUCTION At the start of the war the routine pro- cedures used in scoring tests administered in a number of Navy installations were seriously at fault and resulted in frequent and large errors in the test scores reported. Improvements in the mechanics of test scoring are described in Sec- tion 15.2. The need for speed in constructing the tests of the U. S. Navy Basic Classification Test Bat- tery (Chapter 2) and the tests for selection of officers (Chapter 3) left little time for funda- mental research or development in test con- struction or the theory of mental measurement. It was, however, possible to try out, on a fairly large scale, some ideas which had not been thoroughly tested before the war. Sections 15.3 to 15,5 describe the applications of these ideas. 15 2 IMPROVEMENTS IN THE ADMIN- ISTRATION AND SCORING OF NAVY TESTS Before beginning to develop tests for the Navy Basic Battery or the Officer Qualification Test, detailed analyses were made of the tests in use by the Navy in 1942, This preliminary work included a study of the accuracy with which test papers were being scored at the four naval training stations located at Great Lakes, Illinois; San Diego, California; Newport, Rhode Island; and Norfolk, Virginia.1 Very large errors were found in the scoring of some papers. In the GCT, scoring errors ran as high as 33 points, almost one-third of the total score. For the Code Test of 78 items, errors went up to 40 points, or over half the total score. For the MAT with a total score of 324, 1 paper was in error by 131 points and 8 were in error by more than 50 points. On the Arithmetic, English, and Spelling tests, each with 100 points perfect score, scoring errors of as much as 40 points were found. Such large errors, even though infrequent, are serious when individual assignments are made on the basis of test score. The most accurately scored examination at the most accurate station showed an average error of .06 points on a test containing 29 items. Of the 220 papers checked, 207 were correctly scored. The most inaccurately scored examination at the most inaccurate station showed an average error of 11.41 points on a test on which the perfect score was 342. Of the 204 papers checked, only 6 were correctly scored. The best of these records was three times as a This chapter is based upon the work of Project N-106. RESTRICTED 147 148 CONTRIBUTIONS TO TEST THEORY AND TEST ADMINISTRATION bad as the functioning accuracy level of the College Entrance Examination Board or the Educational Records Bureau (two professional testing organizations). The worst record showed over 50 times as large an average error In order to reduce the number of scoring errors, instructions for a standardized scoring procedure were written. This procedure is out- lined schematically in Figure 1. The packet slip shown in Figure 2 was de- Scanner Scans papers; makes up packets. Preliminary Hand Scorer Scores 10% of papers, recording score on reverse side. Hand Scorer Machine Scorer Score all papers, recording score in answer box. Score Adjuster Compares the two scores, adjusts discrepancies; returns packets with excessive errors for rescoring. (This ends procedure for one-sided answer sheets which DO NOT REQUIRE COMPUTATION.) First Computer Computes total score, recording it on side opposite box for total score. Transcription Clerk Transcribes scores to side with box for total score. Second Computer Computes total score, using data on side with box for total score; records score on the same side. Computation Adjuster Compares total scores, adjusts discrepancies. Figure 1. Suggested scoring procedure for naval training stations. as the maximum ever allowed by these testing organizations. The fact that Army tests were being scored with much greater accuracy than Navy tests indicated that the fault was not a necessary consequence of using enlisted personnel as test scorers. vised as a record of the progress of each packet of answer sheets through the scoring routine and as a means of fixing responsibility for such errors as occurred. Occasional checks made later showed that the adoption of these procedures by the Navy led to a great decrease in the seriousness of scoring RESTRICTED MEASURING THE PREMIUM PLACED ON SPEED IN A TEST 149 errors. Errors continued to be made with some- what greater frequency than would be tolerated by a professional testing organization, but they were fewer in number and smaller in size than those found prior to the introduction of the standard scoring procedures. At the same time, ease of test administration and scheduling requires fixed time limits. If the assigned time limits prove to be too short, what was originally intended to be a power test becomes a speed test. It is desirable, therefore, to have available a method of measuring the Number of papers Packet number Packet to be; Name of test: Hand-scored M.A.T. Machine scored G.C.T. Code A., E., and S. Errors Comments Scanned and counted by Sample of papers hand-scored by Page 1 scored by Page 2 scored by Page 1 adjusted by Page 2 adjusted by First computation of total score by Transcribed by Second computation of total score by Total score adjusted by APPROVED; Selection Officer Figure 2. Packet slip. 15 3 MEASURING THE PREMIUM PLACED ON SPEED IN A TEST Tests which grant a high premium for speed or impose a heavy penalty for slowness are commonly called speed tests. In tests with ex- tremely short time limits, speed is likely to play a major role in determining test scores; in tests with long time limits, speed is likely to have little or no effect upon test scores. For many testing purposes, it is generally considered de- sirable to emphasize power rather than speed. premium which any test, due to its time limits, places on speed of work. The basic assumption in the development of this measure4 was that differences in the speed requirements of tests may be measured by dif- ferences in the proportion of individuals who fail to attempt the final item of each test. The proportion of individuals failing to attempt the final item may be symbolized by S. 5=1.00-^ i RESTRICTED 150 CONTRIBUTIONS TO TEST THEORY AND TEST ADMINISTRATION where Nf = number of persons attempting to answer the final item, and N{ = number of per- sons attempting to answer the initial item, Nf/Ni = the proportion of individuals who attempted the final item; subtracting from 1.00 gives the proportion who did not attempt the final item. S may be thought of either as a measure of the premium for speed or the penalty for slow- more valid, measure would take into account the number of persons attempting to answer each individual item. Comparisons between the values of S for each test and subtest in the U. S. Navy Basic Classification Test Battery and the number of people attempting each item on each subtest indicate that S is a satisfactory measure. Table 1 gives the value of S for each test and subtest in the Basic Battery. The use of S clearly shows differences among these tests, all of which were originally intended primarily as power tests. Deciding whether the value of S is too high, too low, or appropriate for a given test depends, first of all, on whether the test was intended to be a speed test or a power test. For power tests, S should be rather low. There is no very good basis for stating any exact limits. Perhaps the range from .10 to .35 can be suggested as de- sirable for power tests. Judged by this stand- ard, half the tests, or subtests, listed in Table 1 are found to have time limits which are too short. If the items toward the end of a test are quite difficult, it makes little difference in test score whether the slower, and usually poorer, indi- viduals attempt them or not. They miss them even when they have time to answer them. The test can therefore be allotted a shorter time, if the final items are very difficult, than would be desirable in a test where all items are of more nearly equal difficulty. If a test is to be administered by different individuals under different conditions, it is well to err on the side of an excessive time limit (a low value of S) rather than an inadequate one (a high S). The higher the value of S, the more sensitive the test is to irregularities of timing and to unintentional differences among ex- aminers in their emphasis on speed versus accuracy. The use of S is suggested as an easy method of determining the premium placed on speed by any test and as a basis for determining whether or not appropriate time limits have been established. As experience develops, the meaning of smaller differences in S will become clear. It will then become necessary to refine the measure itself. Table 1. Premium for speed, as shown by the value of S* (national sample, iV = 500). Group of items (test or subtest) Ni Nf 5* (Premium for speed) GENERAL CLASSIFICATION TEST: Sentence completion Opposites Analogies 500 500 500 208 364 443 .58 .27 .11 READING 500 310 .38 ARITHMETICAL REASONING 500 220 .56 MECHANICAL APTITUDE TEST: Block counting Mechanical comprehension Surface development 499 500 489 106 338 95 .79 .32 .81 MECHANICAL KNOWLEDGE TEST: Elect.-pictorial Items (Nos. 1-25) Mech.-pictorial items (Nos. 26-55) 500 500 499 421 .00 .16 Elect.-verbal items (Nos. 56-90) Mech.-verbal items (Nos. 91-135) 500 500 379 296 .24 .41 *S is the proportion of individuals who failed to attempt the final item in a test or subtest. Its value is determined from the formula S = 1.00—Nf/Ni in which Nf is the number who attempted the final item and Ni the number who attempted the initial item on the test. The higher S is, the more the test score was influenced by speed. ness. S may range in value from .00 to 1.00— the value .00 indicating the absence of premium for speed, and 1.00 indicating a heavy premium for speed. For example, if time is unlimited, so that everyone completes the test, Nf = S then equals zero (S = 1.00 — 1.00). If the time limit on another test is short, so that only half the men attempt the final item, Nf is only half as large as Nif and S equals .50. The higher the value of S, the greater premium there is on speed, S is based on only two items in a test, the first and the last. A more refined, and probably RESTRICTED SELECTION OF TEST ITEMS BY CORRELATION WITH AN EXTERNAL CRITERION 151 15 4 DETERMINING THE VALIDITY OF A TEST One of the most commonly employed pro- cedures for determining the validity of a test involves the following steps: (1) administering the test to a group of Service school candidates; (2) waiting until the men have completed their Service school training; and then (3) correlat- ing test scores with final achievement grades in the school. Obviously, this procedure demands that one wait as long as three or four months before it becomes possible to make a statement concerning the validity of the test in question. An alternative procedure,3 intended to shorten the time necessary to secure validity data, was tried out in developing the Arith- metical Computation Test (AC) described in Chapter 2. It consisted of the following steps: 1. Administering the test to be validated (AC in this case) to classes at the time of graduation and to other classes just entering training. 2. Correlating AC scores with final school grades for the graduating classes. This gave validity coefficients but not necessarily the same ones which would have been obtained had AC been given before training started, for the training itself might have had a differential effect on the distribution of AC scores. In order to determine whether the obtained coefficients were reasonable estimates of those which would have been found by the usual technique, steps 3, 4, and 5 were carried out. 3. Securing from the Q-cards scores on the Basic Battery tests for both entering and graduating classes. 4. Determining regression coefficients, by the usual formula: b12 = of AC on GCT and AR, of GCT on AR, and of AR on GCT for both entering and graduating classes. 5. Comparing the regressions of AC on GCT and on AR for graduating classes with those for entering classes and comparing the differ- ences obtained with differences in the regres- sion of GCT on AR and AR on GCT. These last two regressions involved tests which both entering and graduating classes had had prior to their entry into school; differential effects of training therefore could have had no influence on the regressions. Variability in these regres- sions could be attributed to sampling differ- ences and could be used as a standard for in- terpreting any differences between entering and graduating class regressions of AC on GCT and on AR. These comparisons were made for nine dif- ferent Service schools. None of the differences between regression coefficients of AC on AR or on GCT for entering and for graduating classes approached statistical significance. It was therefore concluded that the validity co- efficients obtained for the classes tested at the time of graduation were approximately the same as those which would have been obtained for the same classes had they been tested at entry in the usual manner. After the classes to which the AC test was given at the time of entry had graduated, it was possible to use their school grades to vali- date the AC in the usual manner.5 It was thus possible to compare the estimated validity of the AC with the validity as determined by the usual technique. The comparison is shown in Table 2. In only two classes did the two validity coefficients dif- fer by more than .10. The mean of the differ- ences was .001. This difference was smaller than the differences in validity coefficients for most of the Basic Battery tests, for which the time of administration was not a variable. (Data for GCT are given in Table 2; compar- able data for all Basic Battery tests are given in reference 5.) It was concluded that the dif- ferences in AC validity between classes tested at graduation and those tested at entry were due to differences between the samples and not to the influence of training. This method is recommended for further trial in other situations. If it continues to work satisfactorily, it will be of considerable value in speeding up the validation of a new test. is-5 SELECTION OF TEST ITEMS BY CORRELATION WITH AN EXTERNAL CRITERION The criterion usually used in the evaluation of a test item is the total score on the test or RESTRICTED 152 CONTRIBUTIONS TO TEST THEORY AND TEST ADMINISTRATION subtest which includes the item. The criterion here is an internal one. Using an internal cri- terion emphasizes test reliability and tends to secure subtests which are homogeneous in con- tent. It is sometimes desirable to put greater emphasis on test validity than does the usual method of item analysis. In these cases the same external criterion, for example school grades, should be used for item analysis that is used for determining the validity of the total test. A method for conducting this type of item analysis has been suggested by Horst6 and mod- ified by Applied Psychology Panel Project N-106.2 a% = \/piqi, which is the standard deviation of the item; 2?; = the proportion of the population answering an item correctly; (/; = the proportion of the population answering the item incorrectly; rit = the point-biserial correlation of the item with total-test score; and ric = the point-biserial correlation of the item with the external criterion. Note: These five statistics are computed from all cases taking the test and not just from the cases attempting each item. That is, a per- Table 2. Comparisons of validity coefficients of AC test given at time of entry and at time of graduation to kinds of Service schools. men in six N AC GCT Service school Grad. Ent. Grad. Ent. Diff. Grad. Ent. Diff. B. Eng 214 188 .69 .43 - .26 .54 .31 -.23 EM 67 199 .50 .48 -.02 .01 .45 .44 88 77 .41 .59 .18 .45 .55 .10 EC 125 117 .38 .38 .00 .46 .30 - .16 73 67 .39 .29 -.10 .35 .25 -.10 QM 48 137 .33 .39 .06 .21 .46 .25 63 100 .42 .49 .07 .27 .37 .10 SC and Bkr 84 72 .34 .40 .06 .11 .11 .00 SK 115 149 .39 .37 -.02 .38 .44 .06 111 63 .40 .44 .04 .50 .16 -.34 Mean difference .001 .012 Mean absolute difference .081 .178 In attempting to select a set of items such that the correlation with a criterion will be maxi- mized, it is desirable to select individual items which correlate high with the external cri- terion (grades) and relatively low with the in- ternal criterion (total-test score). However, such a statement does not take the difficulty of the item into consideration. If item difficulty is included in a theoretical analysis of this problem, it can be shown that it is desirable to have ricoi for all the selected items as large as possible and ritai as small as possible, where son who records no answer for the last 10 items in a test is counted as having answered those items incorrectly. By this procedure the validity of the test is in- creased at the expense of decreasing its hom- geneity as indicated by the split-half reliability coefficient. The procedure followed was to compute ritoif which is a “reliability” index, and ricGif which is a “validity” index, for each item; to plot these values, with the validity index on the ordinate and the reliability index on the RESTRICTED SELECTION OF TEST ITEMS BY CORRELATION WITH AN EXTERNAL CRITERION 153 abscissa for each of the 104 items in the Me- chanical Comprehension Test—OQT 0-2. Items lying above and to the left of a diagonal line through the origin were then selected for in- clusion in the test. This line was moved down until it included enough items to make up two 30-item tests, called forms 2W and 3W. In order to make the items in the two tests match fairly well with respect to item content, a few items above this line were rejected. The plot is shown in Figure 3. The items selected for ternal criterion; 20 items included by correlat- ing with an internal criterion would not have been chosen by correlating with an external criterion. A sample of the tabulated statistical infor- mation computed for each item and used in selecting items on the basis of an external criterion is given in Table 3. Table 3. Item analysis data against an external criterion. for i items validated Item No. in Form OQT 0-2 Form ric r ica t pq P Nt 39 .10 .20 .0361 .0745 .2346 .62 418 15 2 .53 .43 .0722 .0607 .0668 .93 418 7 2 .47 .20 .1425 .0606 .0177 .77 418 86 2 .29 - .16 .1149 - .0534 .2238 .34 405 Item 39 was an item of approximately aver- age difficulty (p = .62) which did not correlate very well with the other items in the test (rit = .10). It was not included in the final forms of the OQT. Its correlation with indoc- trination school grades (the external criterion) was somewhat higher than its correlation with the other items, so it would have been included had the external criterion of item selection been used. Item 15 was an easy item which showed sub- stantial correlations with both criteria. The use of either criterion would dictate its inclu- sion in the final test. Items 7 and 86 both showed appreciable cor- relations with total-test score and both were included in Form 2 of the OQT. One was some- what below and the other somewhat above average difficulty. When correlated with indoc- trination school grades both showed a low, and one a negative, correlation with the criterion. Both would have been rejected by the external criterion method. Note that items 39 and 7 both had an ric of .20. But one was somewhat more difficult than the other. Item 39, for which ricoi was larger than ritoP was acceptable on the external criterion basis, while item 7, for which ricoi was less than ritGP was not. Figure 3. Item selection plot for the Mechani- cal Comprehension Test—OQT 0-2. form 2W of the test are indicated by circles; the items selected for form 3W by squares; and the rejected items by triangles. The data illustrated in Figure 3 are taken from an analysis of the mechanical comprehen- sion items in the experimental form of the Officer Qualification Test (OQT). Forms 2 and 3 of the OQT were developed from this experi- mental form (see Chapter 3) but not by means of the method being described, that of correla- tion with an external criterion. Instead, the items actually included in Forms 2 and 3 of the OQT were selected by the more usual pro- cedure of correlation with an internal criterion. A comparison of the sets of items selected by the two methods showed much overlapping. Of the 60 items actually included in Forms 2 and 3, 40 were also chosen by correlating with an ex- RESTRICTED 154 CONTRIBUTIONS TO TEST THEORY AND TEST ADMINISTRATION The use of an external criterion for item selection ought to increase the validity of the resulting test over that achieved by using an internal criterion. The available evidence indi- cates that it does. Forms 2 and 3 of the OQT as actually made up, using the customary in- ternal criterion, had validity coefficients of .35 and .38. Validity coefficients of .46 and .48, for the two forms (2W and 3W) in which items were selected on the basis of an external cri- terion, were obtained by rescoring the original test papers using only the 60 items found in this test and computing the correlation of the new scores with indoctrination school grades. These coefficients would no doubt be lowered if the tests were used on a new population, but there is no satisfactory method available for estimating the amount of shrinkage. An em- pirical trial with a new sample was not feasible. Continued study of the value of this method of selecting items by using it in a number of trial situations is strongly recommended. RESTRICTED GLOSSARY LABORATORY NAME ABBREVIATIONS Brown. Brown University, Providence, Rhode Island. C.E.E.B. College Entrance Examination Board, Prince- ton, New Jersey. Harvard. Harvard University, Cambridge, Massa- chusetts. N.A.S. National Academy of Sciences, Washington 25, D. C. N.Y.U. New York University, New York City. Penn. University of Pennsylvania, Philadelphia, Pennsylvania. Penn. State. Pennsylvania State College, State Col- lege, Pennsylvania. Princeton. Princeton University, Princeton, New Jersey. Psych. Corp. Psychological Corporation, 522 Fifth Avenue, New York 18, New York. Rochester. University of Rochester, Rochester, New York. Stanford. Stanford University, California. Tufts. Tufts College, Medford, Massachusetts. U.S.C. University of Southern California, Los Angeles, California. Wis. University of Wisconsin, Madison, Wisconsin. Yerkes Lab. Yerkes Laboratories of Primate Biology, Orange Park, Florida. OTHER ABBREVIATIONS AA. Antiaircraft. AAA. Antiaircraft Artillery. AAF. Army Air Forces or Army Air Field. AC. Arithmetical Computation Test. AGCT. Army General Classification Test. AGO. Adjutant General’s Office. AI. Airborne Interception. APP. Applied Psychology Panel. AR. Arithmetical Reasoning Test. ARC-1. Army Radio Code Aptitude Test (same as SOR). AT. Arithmetic Test. BB. Battleship. BuPers. Bureau of Naval Personnel. CHL. Chain Home Link. CIC. Combat Information Center. COMINCH. Commander in Chief, United States Navy. COTCLANT. Commander, Operational Training Com- mand, United States Atlantic Fleet. COTCPAC. Commander, Operational Training Com- mand, United States Pacific Fleet. CRT. Code Receiving Tests. DD. Destroyer. DE. Destroyer Escort. ESF. Equivalent square feet. EW. Early Warning. FC. Fire Controlman. Form 20. War Department enlisted men’s personal record card. GCI. Ground Controlled Interception. GCT. General Classification Test (Navy). JL. A shipboard telephone circuit. LCVP. Landing Craft, Vehicles and Personnel, MAT. Mechanical Aptitude Test. MK-E. Mechanical Knowledge, Electrical. MK-M. Mechanical Knowledge, Mechanical. N. Number. Nf. The number of men in a group who attempted to answer a particular test item. NavPers. Bureau of Naval Personnel publication number. NDRC. National Defense Research Committee. NRC. National Research Council. NROTC. Naval Reserve Officers’ Training Corps. NTS, Naval Training Station. NTSch. Naval Training School. OCS. Officer Candidate School. OQT. Officer Qualification Test. OSRD. Office of Scientific Research and Development. PI. Personal Inventory. PPL Plan Position Indicator. Q-Card. Navy Enlisted Men’s Qualification Card. R. Reading Test. r. Correlation coefficient. rl'ls. Biserial correlation coefficient. S. A measure of the premium placed on speed by a test. a. Standard deviation, SCCAT. Signal Corps Code Aptitude Test. S.D. Standard deviation. SOR. Speed of Response Test of Code Aptitude. UOE. Unit of Error. V-12. A Navy college training program. WD. War Department. RESTRICTED 155 BIBLIOGRAPHY CHAPTER 1 1. The Use of Test Scores and Quality-Classification Ratings in Predicting Success in Electrician’s Mates Schools, Herbert S. Conrad and G. A. Satter, OSRD 5667, CEEB,* Sept. 13, 1945. APP-420-M2 2. Report of Observers, ETO, 11 March to 21 April 19U5, Willard Machle and Frederick S. Brackett, Headquarters ETO, U. S. Army, APO 887, to Office of the Surgeon General, Washington, D. C., Apr. 27, 1945. CHAPTER 2 1. Recommendations on Selection, Classification, and Training of Service Personnel, Memorandum to Chief of the Bureau of Naval Personnel, Washing- ton, D. C., from the Committee on Service Per- sonnel—Selection and Training, National Research Council, July 25, 1942. 2. Naval Aptitude Tests: A. Reliability; B. Scoring Accuracy; C. Scoring Procedures, Harold Gulliksen, OSRD 1127, CEEB, Dec. 10, 1942. APP-322-M1 3. Preparation of the United States Navy General Classification Test—Form 1, and the United States Navy Tests of Reading and Arithmetical Reason- ing, Form 1, H. Gulliksen, Project N-106, Memo- randum 7, CEEB, June 23, 1943. APP-327-M1 4. Averages, Standard Deviations, and Intercorrela- tions of Navy Aptitude Tests, Harold Gulliksen, Herbert S. Conrad, and Norman Frederiksen, OSRD 1536, CEEB, June 7, 1943. APP-322-M2 5. Suggestions for the Revision of the United States Navy Mechanical Aptitude Test, Form T, H. Gulliksen, Project N-106, Memorandum 2, Jan, 18, 1943. APP-325-M1 6. Factor Analysis of the New United States Navy Basic Classification Test Battery, D. A. Peterson, OSRD 3004, CEEB, Sept. 29, 1943. APP-412-M1 7. Item Analysis of Navy Aptitude Tests, Herbert S. Conrad, OSRD 3039, CEEB, Dec. 30, 1943. APP-322-M3 8. Validity of Navy Aptitude Tests in Service Schools at the Great Lakes Naval Training Station, Norman Frederiksen, OSRD 3245, CEEB, Jan. 31, 1944. APP-322-M4 9. Statistical Analysis of the Mechanical Knowledge Test, Herbert S. Conrad, OSRD 3246, CEEB, Jan. 28, 1944. APP-325-M3 10. Validity of an Experimental Battery of Aptitude Tests at the Ordnance and Gunnery Schools, Wash- ington Navy Yard, Norman Frederiksen, OSRD 3619, CEEB, Apr. 29, 1944. APP-322-M5 11. Selection of Items for the US Navy General Classi- fication Test, Form 2, and the US Navy Tests of Reading and Arithmetical Reasoning, Form 2, G. A. Satter, OSRD 3756, CEEB, June 8, 1944. APP-412-M3 12. The Preparation of Norms for the Fleet Edition of the General Classification Test, D. A. Peterson, OSRD 4242B, CEEB, Oct. 10, 1944. APP-414-M1 13. The Construction and Validation of an Arithmetical Computation Test, G. A. Satter and Norman Frederiksen, OSRD 4556, CEEB, Jan. 8, 1945. APP-411-M2 14. A Statistical Evaluation of the Basic Classification Test Battery, Form 1, Herbert S. Conrad, OSRD 4636, CEEB, May 14, 1945. APP-411-M3 15. A Further Study of the Validity of the Arithmetical Computation Test, Norman Frederiksen, OSRD 5302, CEEB, July 3, 1945. APP-411-M4 16. Summary Report on Research and Development of the Navy’s Aptitude Testing Program, Herbert S. Conrad, OSRD 6110, CEEB, Oct. 31, 1945. APP-322-M6 17. Statistical Method, Truman L. Kelley, Macmillan Co., New York, N. Y., 1923, Formula 186, p. 225. 18. Enlisted Classification Bulletin, Navy Bureau of Personnel, September 1945, p. 8. 19. Predictive Efficiency of the Navy Basic Test Battery at Naval Training School (Gunners Mates —Electrical Hydraulics), Navy Yard, Washington, D. C., Project 224, Navy Bureau of Personnel, Test and Research Section, July 1, 1945. 20. “Psychological Test Construction and Research in the Bureau of Naval Personnel: Development of the Basic Test Battery for Enlisted Personnel,” Staff of the Test and Research Section, Psycholog- ical Bulletin, Vol. 42, 1945, pp. 561-571. 21. “Psychological Test Construction and Research in the Bureau of Naval Personnel: Validity of the Basic Test Battery, Form 1, for Selection for Ten Types of Elementary Naval Training Schools,” Staff of the Test and Research Section, Psycho- logical Bulletin, Vol. 42, 1945, pp. 638-644. a College Entrance Examination Board. RESTRICTED 157 158 BIBLIOGRAPHY CHAPTER 3 1. Preparation of a Preliminary Form of an Officers’ Selection Test, H. Gulliksen, Project N-106, Memo- randum 1, CEEB, Dec. 11, 1942. APP-319-M1 2. Analysis of the N.R.O.T.C. Selective Examination (Form C) and Suggestions for Its Revision, Harold Gulliksen, OSRU 1290, CEEB, Feb. 15, 1943. APP-411-M1 3. Preparation of the United States Navy Officer Qualification Test, Form 1, Harold Gulliksen, OSRD 1273, CEEB, Mar. 12, 1943. APP-319-M2 4. Preparation of Norms for the Officer Qualification Test, Form 1, Norman Frederiksen, Project N-106, Memorandum 4, CEEB, May 28, 1943. APP-319-M3 5. Comparison of the Officer Qualification Test, Form 1, and the United States Navy Aptitude Test, Form E-2, Norman Frederiksen, Project N-106, Memo- randum 5, CEEB, June 1, 1943, 6. Preparation of Norms for Women for the Officer Qualification Test, Form 1, H. Gulliksen, Project N-106, Memorandum 9, CEEB, July 24, 1943. APP-319-M4 7. Development and Validity of the United States Navy Officer Qualification Test, Norman Frederik- sen and D. A. Peterson, OSRD 3186, CEEB, Jan. 7, 1944. APP-319-M5 8. Study of Combat Leadership, W. S. Hunter, Memo- randum to General Liaison Officers of the Applied Psychology Panel, NDRC, Mar. 3, 1944. 9. Report to Applied Psychology Panel on Combat Leadership, Henry E. Garrett and Ernest M. Ligon, NAS, June 8, 1944. 10. A Statistical Evaluation of the United States Navy Officers Qualification Test, Forms 2 and 3, D. A. Peterson, OSRD 3978, CEEB, Aug. 4, 1944. APP-319-M6 11, Preliminary Summary of Results of the Study of the Effectiveness of the Classification Procedure for Officers, Amphibious Training Command, US Atlantic Fleet, Kinsley R. Smith and Harold A. Voss, Project N-117, Informal Memorandum 6, Pennsylvania State College, Apr. 12, 1945. APP-512-M5 CHAPTER 4 1. Standardization and Validation of the Personal In- ventory : Psychiatric Criterion, W. C. Shipley, Florence E. Gray, and Nancy Newbert, OSRD 1606, Brown University, June 24, 1943. APP-321-M1 2. A Report of Research on Selection Tests at the US Submarine Base, New London, Connecticut, C. H. Graham, L. A. Riggs, N. R. Bartlett, R. N. Berry, W. S. Verplanck, R. L. Solomon, and C. G. Mueller, OSRD 1770, Brown University, June 28, 1943. 3. Results Obtained from Testing Recruits with the New London-NDRC Questionnaire at the Newport Naval Training Station, F. A. Mote, R. N. Berry, and C. H. Graham, OSRD 3040, Brown University, Dec. 6, 1943. APP-324-M1 4. The Relation of Selection Test Scores to Tank Escape Performance at the Submarine School, C. H. Graham, F. A. Mote, and R. N. Berry, OEMsr-834, OSRD 3262, Project N-113, Brown University, Jan. 31, 1944. APP-413-M2 5. Item Analysis and Evaluation of the Scoring Stencil of the Personal Inventory, W. C. Shipley, Florence E. Gray, and Nancy Newbert, OSRD 3315, Brown University, Feb. 14, 1944. APP-412-M2 6. Test Scores of Amphibious Forces Personnel at Solomons, Maryland, Contrasted with Those Ob- tained by Personnel of Other Navy Activities, C. H. Graham, Project N-113, Memorandum 1, Brown University, Mar. 14, 1944. 7. The Personal Inventory, Short Form, Format C: Derivation and Preliminary Psychiatric Validation, W. C. Shipley, Florence E. Gray, and Nancy New- bert, OSRD 3390, Brown University, Mar. 15, 1944. APP-321-M3 8. The Comparability of Formats A and B of the Personal Inventory, R, N. Berry, H. J. Leavitt, and F. A. Mote, OSRD 3582, Brown University, Apr. 21, 1944. APP-321-M4 9. The Personal Inventory, Short Form, Format C: Psychiatric Validation on a Pre-Testing Basis, W. C. Shipley, Florence E. Gray, and Nancy New- bert, OSRD 3604, Brown University, May 1, 1944. APP-321-M5 10. A Comparison of Personal Inventory Scores with Service Records One Year After Testing, Walter C. Shipley, Florence E, Gray, and Nancy Newbert, OSRD 3755, Brown University, June 10, 1944. APP-321-M6 11. Results From the Long and Short Forms of the Personal Inventory and the General Classification Test, R. N. Berry, C. H. Graham, and F. A. Mote, OSRD 3962, Brown University, July 31, 1944. APP-321-M7 12. Final Report in Summary of Research on the Per- sonal Inventory and Other Tests, W. C. Shipley and C. H. Graham, OSRD 3963, Brown University, Aug. 1, 1944. APP-321-M8 RESTRICTED BIBLIOGRAPHY 159 13. Item Differentiation and Derivation of a 50-Item Scoring Key for the Officer Personal Inventory, Form 1, Kinsley R. Smith and Harold A. Voss, Project N-117, Memorandum 4, Pennsylvania State College, Feb. 8, 1945. APP-412-M5 14. The Selection of Emotionally Unstable Personnel by Battle Noise Equipment, Eugene L. Hartley and Dorothea B. Jones, OSRD 4684, Brown Uni- versity, Feb. 10, 1945. APP-160-M1 15. An Evaluation of the Personal Inventory for Pre- dicting Success in Parachute School, G. A. Satter, OSRD 4870, CEEB, Mar. 28, 1945. APP-321-M9 16. NDRC Battle Noise Equipment, Utilization at In- filtration Range, Eugene L. Hartley, Project N-113, Informal Memorandum 6, Brown University, Apr. 3, 1945. APP-630-M2 17. Final Summary of Research on the Use of Battle Noise Equipment, Eugene L. Hartley and Dorothea B. Jones, OSRD 4931, Brown University, Apr. 12, 1945. APP-630-M3 18. An Evaluation of the Personal Inventory and Cer- tain Other Measures in the Prediction of Submarine Officers’ Evaluations of Enlisted Men, G. A. Satter, OSRD 5557, CEEB, Sept. 7, 1945. APP-321-M10 19. Intercorrelations of Items of the NDRC Personal Inventory, and Correlations of These Items with Otis Intelligence Scores, N. R. Bartlett, Bureau of Medicine and Surgery Research Project X-460 (Sub. 112), Medical Research Department, U. S. Submarine Base, New London, Conn., June 2, 1945. 20. Application of the NDRC Personal Inventory to Marine Corps Officer Candidates: Preliminary Validation, S. B. Williams and H. J. Leavitt, Bureau of Medicine and Surgery Research Project X-386 (Sub. 75), Camp LeJeune, N. C., Dec. 26, 1944. 21. Application of the NDRC Personal Inventory to Marine Corps Officer Candidates, Progress Report No. 2: Further Validation, S. B. Williams, H. J. Leavitt, and G. V. Mendola, Bureau of Medicine and Surgery Research Project X-386 (Sub. 75), Medical Field Research Laboratory, Camp LeJeune, N. C., Nov. 27, 1945. 22. Application of the NDRC Personal Inventory to Marine Corps Officer Candidates, Progress Report No. 3: Data on Final 671 Candidates, H. J. Leavitt, S. B. Williams, and N. J. Lipkin, Bureau of Medi- cine and Surgery Research Project X-386 (Sub. 75), Medical Field Research Laboratory, Camp LeJeune, N. C., Dec. 20, 1945. 23. The Development of the MFRL Confidential Ques- tionnaire, S. B. Williams, H. J. Leavitt, and C. L. Blair, Bureau of Medicine and Surgery Research Project X-615 (Sub. 133), Medical Field Research Laboratory, Camp LeJeune, N. C,, Dec. 21, 1945. 24. History of AAF Aviation Psychology Program. (Specific information not available.) CHAPTER 5 1. Report on an Activity-Preference Test for the Classification of Service■ Personnel, Truman L. Kelley, OSRD 4484, Final Report of Project SOS-7, Harvard University, Dec. 21, 1944. APP-521-M2 CHAPTER 6 1. The Experimental Edition of Code Receiving Tests, Albert K. Kurtz, Harold G. Seashore, and John M. Willits, OSRD 1314, Psychological Corp., Mar. 29, 1943. APP-317-M1 2. Instructor’s Manual for the Code Receiving Tests, NDRC Project N-107, Psychological Corp., Novem- ber 1943. APP-317-M2 3. Tentative Manual and Guide for Using the Radio Code Test: Speed of Response, Form 1, NDRC Project N-107, Psychological Corp., Dec. 15, 1943. APP-317-M3 4. The Prediction of Code Learning Ability, Albert K. Kurtz, OSRD 4059, Psychological Corp., Aug. 26, 1944. APP-212.1-M1 5. Experiments in Training Radar Operators in Visual Code Reception, Irving H. Anderson, Alfred L. Baldwin, Robert S. Daniel, William H. Lichte, Thomas L. McCulloch, Fred McKinney, and Donald B. Lindsley, OSRD 4811, Yerkes Laboratories, Mar. 20, 1945. APP-212.5-M1 6. “Fine Eignungsprufung fur Funkentelegraphis- ten,” R. A. Biegel, Psychotechnische Zeitschrift, Vol. 6, April 1931, pp. 41-45. 7. Report on a Code Aptitude Test, L. L. Thurstone, University of Chicago, May 1943. 8. Report on Radio Code Aptitude Tests, PR U000, Adjutant General’s Office, Technical Section, Classi- fication and Replacement Branch, 1944. 9. Memorandum on Comparison of Code Aptitude Test (Thurstone) and Radio Code Aptitude Test (Speed of Response), Pers-4122b-do, U. S. Navy, June 7, 1944. CHAPTER 7 1. Instruction Manual for Oscilloscope Operator Tests: Experimental Edition, Donald B. Lindsley, OSRD 1811, Yerkes Laboratories, Sept. 1, 1943. APP-318-M2 2. Oscilloscope Operator Tests: Experimental Edi- tion, Donald B. Lindsley, OSRD 1812, Yerkes Lab- oratories, Sept. 1, 1943. APP-318-M1 RESTRICTED 160 BIBLIOGRAPHY 3. Preliminary Report of Results from Oscilloscope Operator Tests: Experimental Edition, Donald B. Lindsley, OSRD 1813, Yerkes Laboratories, Sept. 1, 1943. APP-318-M3 4. Proficiency Measures for Classification and Assign- ment of Operators of ASV Radar Equipment, Irving H. Anderson and others, OSRD 2012, Yerkes Laboratories, Sept. 29, 1943. APP-318.1-M1 5. Development of Proficiency Measures for Classifica- tion and Assignment of Operators of ANU Radar Equipment, Irving H. Anderson and others, OSRD 2013, Yerkes Laboratories, Sept. 29, 1943. APP-318.1-M2 6. Vision as Related to Proficiency in Oscilloscope Operation, Irving H. Anderson and others, OSRD 3409, Yerkes Laboratories, Feb. 24, 1944. APP-318-M5 7. A Validational Study of Oscilloscope Operator Tests, Irving H. Anderson and others, OSRD 3712, Yerkes Laboratories, Apr. 24, 1944. APP-318-M6 8. Report on Selection of Intercept Officer Candidates, George K. Bennett, NDRC Project N-109, Informal Memorandum 3, Psychological Corp,, Jan. 19, 1945. APP-170-M2 9. Prediction of Success in Training and Measurement of Achievement at Naval Training School (Tactical Radar), Bureau of Naval Personnel Project 412 Research Report, May 1, 1945. CHAPTER 8 1. Selection of Stereoscopic Range Finder Operators, OSRD 410, Section D-2, Fire Control, NDRC, Progress Report to the Services 22, Feb. 23, 1942. Div. 7-220.14-M16 2. Description of Instruments for Use in the Selection of Stereoscopic Range Finder Operators, OSRD 592, Section D-2, Fire Control, NDRC, Report to the Services 30, May 18, 1942. Div. 7-220.14-M4 3. Manual for Use in the Selection of Stereoscopic Range Finder Operators, OSRD 721, Section D-2, Fire Control, NDRC, Report to the Services 31, July 7, 1942. Div. 7-220.14-M11 4. Analysis of Factors Effective in the Use of the Stereoscopic Range and Height Finder, OSRD 862, Section D-2, Fire Control, NDRC, Report to the Services 32, Sept. 9, 1942. Div. 7-220.213-M3 5. The Chromatic Dispersion of the Human Eye and Its Possible Influence on Stereoscopic Range Find- ing, OSRD 923, Section D-2, Fire Control, NDRC, Report to the Services 35, Oct, 1, 1942. Div. 7-230.1-M2 6. Supplement to: Manual for Use in the Selection of Stereoscopic Range Finder Operators, OSRD 1210, Div. 7, Fire Control, NDRC, Report to the Services 49, Feb. 17, 1943. Div. 7-220.14-M15 7. Sex Differences in Stereoscopic Ranging and in Visual Tracking, OSRD 1250, Div. 7, Fire Control, NDRC, Report to the Services 52, Mar. 9, 1943. Div. 7-220.34-M2 8. Manual for the Adjustment and Operation of the Projection Eikonometer, H. A. Imus, OSRD 1340, Brown University, Mar. 29, 1943. APP-122-M1 9. Distribution of Measures of Interpupillary Dis- tance, H. A. Imus, OSRD 1341, Brown University, Mar. 24, 1943. APP-121-M1 10. The Adjustment of the M2 Trainer for Standard Testing, H. M. Fisher and H. A. Imus, OSRD 1638, Brown University, June 17, 1943. APP-652-M3 11. The Relationship between Test Scores Obtained on the Single and Multiple Projection Eikonometers, H. A. Imus, OSRD 1789, Brown University, Aug. 5, 1943. APP-122-M2 12. The Selection of Stereoscopic Height Finder Ob- servers, W. C. Biel, W. J. Brogden, W. F. Dear- born, H. M. Fowler, C. H. Graham, H. A. Imus, W. E. Kappauf, and F. A. Mote, OSRD 1790, Brown University, Aug. 14, 1943. APP-120-M1 13. The Relationship of Visual Acuity to Acuity of Stereoscopic Vision, J. K, Adams, H. M. Fowler, and H. A. Imus, OSRD 2087, Brown University, Sept. 15, 1943. APP-121-M2 14. The Reliability and Precision of the NDRC and Bausch and Lomb Interpupillometers, J. K. Adams, D. C. Beier, and H. A. Imus, OSRD 3475, Univer- sity of Wisconsin, Mar. 29, 1944. APP-112.2-M1 15. Final Report in Summary of Work on the Selection and Training of Stereoscopic Height Finder Ob- servers, W. C. Biel, W. J. Brogden, C. H. Graham, H. A. Imus, and W. E. Kappauf, OSRD 3773, Brown University, June 14, 1944. APP-120-M2 16. The Reliability of Two Forms of an Activity Pref- erence Blank, J. K. Adams and H. M. Fowler, OSRD 3778, University of Wisconsin, June 19, 1944. APP-112.1-M1 17. The Construction and Calibration of the NDRC Interpupillometer, D. G. Ellson and H. A. Imus, OSRD 3797, University of Wisconsin, June 19, 1944. APP-112.2-M2 18. A Test-Retest Reliability Study of the Bausch and Lomb Ortho-Rater with Naval Personnel, J. K. Adams, D. C. Beier, and H. A. Imus, OSRD 3969, University of Wisconsin, Aug. 1, 1944. APP-111.1-M1 RESTRICTED BIBLIOGRAPHY 161 19. Manual for Use in the Selection of Fire Controlmen (0), (Stereoscopic Rangefinder Operators), H. A. Imus, OSRD 4050, University of Wisconsin, Aug. 22, 1944. APP-111-M1 20. Manual for the Installation and Adjustment of the Multiple Projection Eikonometer, H. A. Imus, OSRD 4233, University of Wisconsin, Oct. 10, 1944. APP-122-M3 21. A Folloiv-up Study of the Efficiency of the Projec- tion Eikonometer Test in Predicting the Perform- ance of Stereoscopic Height Finder Observers, D. C. Beier, Florence Gray, H. A. Imus, and E. B. Knauft, OSRD 4352, University of Wisconsin, Nov. 21, 1944. APP-122-M4 22. The Selection of Fire Controlmen (0)—Rangefinder and Radar Operators, D, C. Beier and Florence Gray, OSRD 4861, University of Wisconsin, Mar. 26, 1945. APP-111-M2 23. Final Report in Summary of Research on the Selec- tion and Training of Rangefinder and Radar Opera- tors, W. J. Brogden, OSRD 5358, University of Wisconsin, July 23, 1945. APP-110-M1 24. Intercorrelations between Scores on the Tufts Stereo-Trainer; the Navy Stereo-Trainer; the Bott Test of Stereo Acuity; the Vectographic Pursuit Apparatus; the Wulfeck Group Test of Stereo Acuity; and Two Tests of General Intelligence, Project 10, Progress Report to NDRC, Tufts Col- lege, June 18, 1942. Div. 7-220.15-M3 25. The Correlation between the Keystone and Bott Tests of Stereopsis and Influence of Size of Test Stimuli in Judgments of Stereopsis, Project 10, Progress Report to NDRC, Tufts College, June 18, 1942. Div. 7-220.15-M6 26. The Test-Retest Reliability Coefficients of the Tufts Stereo-Trainer, the Navy Stereo-Trainer, Mark II, and the Bott Test of Stereo Acuity, Project 10, Progress Report to NDRC, Tufts College, June 18, 1942. Div. 7-220.15-M4 27. Validation and Standardization of Tests Used in the Selection of Stereoscopic Rangefinder Operators, Fire Control Research, Progress Report 15 to NDRC, Princeton University, July 13, 1942. 28. Interrelationships among Seven Tests of Stereo- scopic Acuity and the Relationship between Two Tests of Visual Acuity and Two Tests of Phorias, H. M. Fowler, H. A. Imus, and F. A. Mote, Memo- randum 12, Brown University, Mar. 24, 1944. APP-121-M3 29. GCT and MAT Selection Standards for Fire Con- trolmen (R), W. J. Brogden, Project N-114, In- formal Memorandum 1, Apr, 28, 1944. APP-112-M1 CHAPTER 9 1. How to Use the NDRC Model III Dark Adap- tometer, Rochester, N. Y., December 1942. APP-151-M1 2. A Study of the Prediction of Night Lookout Per- formance, Carl H. Wedell, OSRD 3357, Princeton University, Mar. 15, 1944. APP-150-M1 3. General Recommendations Concerning Lookouts, Carl H. Wedell, Memorandum to the Applied Psychology Panel, Project N-115, Memorandum 1, Princeton University, Sept. 15, 1944. APP-150-M2 4. A Study of the Perfouvnance of Night Lookouts Aboard Ship, W. C. H. Prentice, OSRD 4087, Princeton University, Oct. 7, 1944. APP-150-M3 5. Final Report in Summary of Work on the Selection and Training of Night Lookouts, C. H. Wedell, OSRD 4342, Princeton University, Nov. 15, 1944. APP-150-M5 CHAPTER 10 1. A Speech Interview for the Selection of Telephone Talkers, George K. Bennett and John W. Black, OSRD 1769, Psychological Corp., August 1943. APP-130-M1 2. A Preliminary Study of the Abilities of Rated Men to Judge Speaking Performance, John C. Snidecor, OSRD 1823, Psychological Corp., August 1943. APP-218.31-M1 3. Report on Seminar on Telephone Talking and Com- position of a Manual for the Use of Sound-Powered Phones, John C. Snidecor, NDRC Project N-109, Informal Memorandum 1, Psychological Corp., Sept. 28, 1943. APP-218.3-M1 4. A Manual for the Training of Interviewers, John C. Snidecor and Grant Fairbanks, OSRD 1827, Psychological Corp. APP-218.31-M3 5. A Study in Training Classification Petty Officers to Select Telephone Talkers, G. K. Bennett and J. W. Black, OSRD 1931, Psychological Corp., November 1943. APP-218.31-M2 6. Methods of Training Telephone Talkers for In- creased Intelligibility, John C. Snidecor, Louis A. Mallory, and Edward L. Hearsey, OSRD 3178, Psychological Corp., Jan. 28, 1944. APP-218.2-M1 7. Information Regarding an Analysis of Recordings Made from the Telephone Circuits on Board a Battleship and a Destroyer, G. K. Bennett and J. W. Black, OSRD 3312, Psychological Corp., February 1944. APP-218.3-M2 8. Final Report in Summary of Work on the Selection and Training of Telephone Talkers, Louis A. Mal- lory and William J. Temple, OSRD 5497, Psycho- logical Corp., Aug. 27, 1945. APP-218.1-M2 RESTRICTED 162 BIBLIOGRAPHY CHAPTER 11 1. Manual of Instruction for the Use of Findex Equip- ment in Personnel Placement: USS New Jersey, John M. Stalnaker, Committee on Service Per- sonnel—Selection and Training, National Research Council, June 28, 1943. APP-521-M1 2. An Analysis of Men Entering the DD Pool, Pre- commissioning Training Center, Treasure Island, May 29 to July 29, 19UU, Max M. Levin, OSRD 4151, Stanford University, Sept. 20, 1944. APP-513-M1 3. A Procedure for Sorting the Enlisted Personnel Qualifications Cards, G. A. Satter and N. C. Kep- hart, OSRD 4689, CEEB, Jan. 22, 1945. APP-532-M1 4. The Selectometer, a Classification Guide, Evan R. Keisler, OSRD 4746, Stanford University, Feb. 24, 1945. APP-521-M3 5. Comments on the Manual Entitled “Handbook of Interviewing Guides for Destroyer (2200-Ton) Billets, Repair Party, Engine Room, Fire Room” (with accompanying handbook), Morris S. Viteles and John H. Gorsuch, NDRC Project NR-106, Memorandum 11, University of Pennsylvania, May 14, 1945. APP-511-M2 6. Classification Data Available to Ships’ Officers, D. A. Peterson, OSRD 5145, CEEB, May 30, 1945. APP-180-M1 7. A Point Score Method for Evaluating Naval Per- sonnel, Max M. Levin, Ronald K. Campbell, Evan R. Keisler, and J. Bradley Sonderman, OSRD 5197, Stanford University, June 12, 1945. APP-522-M1 8. An Experimental Personnel Record System for Shipboard Use, OSRD 5303, CEEB, July 3, 1945. APP-531-M2 9. Memorandum on the Construction and Validation of a Work Readiness Test for Distilling Plant Operators, B. J. Covner, H. A. Voss, and S. M. Wesley, OSRD 5507, University of Pennsylvania, Aug. 29, 1945. APP-316-M1 10. Memorandum on Constructing Work Readiness Tests, H. A. Voss and S. M. Wesley, OSRD 5521, University of Pennsylvania, Aug. 31, 1945. APP-328-M1 11. Final Report in Summary of Research and Devel- opment of Classification Aids by NDRC Project N-116a, Ronald K. Campbell, OSRD 5647, Stanford University, Sept. 11, 1945. APP-520-M1 12. Handbook of Interviewing Guides for Billets on 2,200-Ton Destroyers: Ship Control, Combat Infor- mation, Communication, Damage Control, Engineer- ing, Gunnery (with accompanying handbook), A. G. Bayroff and R. M. Jones, OSRD 6115, University of Pennsylvania, Oct. 16, 1945. APP-511-M4 13, Manual of Classification Procedures for Amphibious Training Bases, Amphibious Training Command, U. S. Atlantic Fleet, Naval Operating Base, Nor- folk, Va., Sept. 1, 1944. CHAPTER 12 1. A Point Score Method for Evaluating Naval Per- sonnel, Max M. Levin, Ronald K. Campbell, Evan R. Keisler, and J. Bradley Sonderman, OSRD 5197, Stanford University, June 12, 1945. APP-522-M1 2. Predicting Success in Service School from the Order of Assignment, G. A. Satter and Herbert S. Conrad, OSRD 5556, CEEB, Sept. 7, 1945. APP-420-M1 3. Use of Test Scores and Quality-Classification Rat- ings in Predicting Success in Electrician’s Mates School, Herbert S. Conrad and G. A. Satter, OSRD 5667, CEEB, Sept. 13, 1945. APP-420-M2 4. Summary Report on Research and Development of the Navy’s Aptitude Testing Program, Herbert S. Conrad, OSRD 6110, CEEB, Oct. 31, 1945. APP-322-M6 5. Manual for Processing the Enlisted Personnel Qualifications Card, U. S. Navy, NavPers 16704, January 1945. 6. Manual of Procedures for U. S. Naval Training Station Selection Department, U. S. Navy, NavPers 16710. CHAPTER 13 1. A Speech Interview for the Selection of Telephone Talkers, George K. Bennett and John W. Black, OSRD 1769, Psychological Corp., August 1943. APP-130-M1 2. Proficiency Measures for Classification and Assign- ment of Operators of ASU Radar Equipment, Irv- ing H. Anderson and others, OSRD 2012, Yerkes Laboratories, Sept. 29, 1943. APP-318.1-M1 3. Memorandum on Service School Grades, Harold Gulliksen, OSRD 3177, CEEB, Jan. 6, 1944. APP-413-M1 4. Manual for Use in the Selection of Fire Controlmen (0), (Stereoscopic Rangefinder Operators), H. A. Imus, OSRD 4050, University of Wisconsin, Aug. 22, 1944. APP-111-M1 5. The Code Receiving Tests, Albert K. Kurtz and Harold Seashore, OSRD 3157, Psychological Corp., Sept. 12, 1944. APP-317-M4 6. A Procedure for Sorting the Enlisted Personnel Qualifications Cards, G. A. Satter and N, C. Kep- hart, OSRD 4689, CEEB, Jan. 22, 1945. APP-532-M1 RESTRICTED BIBLIOGRAPHY 163 7. Final Report in Summary of Work on the Selection and Training of Cargo Handling Teams for Combat Laden Vessels, Floyd L. Ruch, OSRD 5140, USC, May 29, 1945. APP-222-M1 8. Classification Data Available to Ships’ Officers, D. A. Peterson, OSRD 5145, CEEB, May 30, 1945. APP-180-M1 9. The Development of Achievement Tests for Gun- ner’s Mates Schools, Norman Frederiksen and A. E. Monroe, OSRD 5259, CEEB, June 25, 1945. APP-311-M1 10. An Experimental Personnel Record System for Shipboard Use, OSRD 5303, CEEB, July 3, 1945. APP-531-M2 11. Final Report in Summary of Work on the Job Analysis, Qualification, and Placement of Personnel in the Amphibious Force, Kinsley R. Smith, OSRD 5422, Pennsylvania State College, Aug. 8, 1945. APP-512-M8 12. Achievement Examinations for Signalman School, M. Duane Bown, Thelburn L. Engle, and Harold P. Bechtoldt, OSRD 5460, CEEB, Aug. 20, 1945. APP-312-M1 13. Memorandum on the Construction and Validation of a Work Readiness Test for Distilling Plant Oper- ators, B. J. Covner, H. A. Voss, and S. M. Wesley, OSRD 5507, University of Pennsylvania, Aug. 29, 1945. APP-316-M1 14. Development of Achievement Tests for Class “A” Naval Training Schools (Torpedomen), Charles M. Harsh and Joseph B. Cooper, OSRD 5520, CEEB, Aug. 31, 1945. APP-311-M2 15. Memorandum on Constructing Work Readiness Tests, H. A. Voss and S. M. Wesley, OSRD 5521, University of Pennsylvania, Aug. 31, 1945. APP-328-M1 16. An Evaluation of the Personal Inventory and Cer- tain Other Measures in the Prediction of Submarine Officers’ Evaluations of Enlisted Men, G. A. Satter, OSRD 5557, CEEB, Sept. 7, 1945. APP-321-M10 17. The Development of Performance Tests for Use in Class “A” Electrical Schools, Merrell E. Thompson, and Norman Frederiksen, OSRD 5666, CEEB, Sept. 13, 1945. APP-315-M1 18. Development of Achievement Tests for the Landing Craft School, Coronado, California, Norman Fred- eriksen, OSRD 5670, CEEB, Sept. 14, 1945. APP-314-M1 19. Development of Radio Code Receiving Examina- tions, M. Duane Bown and Harold O. Gulliksen, NDRC Project N-106, Memorandum 18, CEEB, Sept. 27, 1945. APP-317-M7 20. Summary Report on Research and Development of the Navy’s Aptitude Testing Program, Herbert S. Conrad, OSRD 6110, CEEB, Oct. 31, 1945. APP-322-M6 21. Manual of Procedures for U. S. Naval Classification Centers, NavPers 15082, Classification and Selec- tion Section, Bureau of Naval Personnel, February 1945. 22. Are Destroyer Radiomen Ready for Battle? [U. S. Navy publication. (?)] 23. Memorandum to Director of Planning and Control, J. B. Hogle, Pers 6370-HMG-l, Bureau of Naval Personnel, Jan. 3, 1946. 24. Manual of Navy Enlisted Job Classification, Nav- Pers 15105, U. S. Navy. CHAPTER 14 1. Naval Aptitude Tests: A. Reliability; B. Scoring Accuracy; C. Scoring Procedures, OSRD 1127, CEEB, Dec. 10, 1942. APP-322-M1 2. Preparation of a Preliminary Form of an Officers’ Selection Test, NDRC Project N-106, Memorandum 1, CEEB, Dec. 11, 1942. APP-319-M1 3. Suggestions for the Revision of the United States Navy Mechanical Aptitude Test, Form T, NDRC Project N-106, Memorandum 2, CEEB, Jan. 18, 1943. APP-325-M1 4. Analysis of the NROTC Selective Examination, Form C, and Suggestions for Its Revision, OSRD 1290, CEEB, Feb. 15, 1943. APP-411-M1 5. Preparation of the United States Navy Officer Qualification Test, Fomn 1, OSRD 1273, CEEB, Mar. 12, 1943. APP-319-M2 6. Preparation of an Experimental Form of a Spatial Relations Test, Norman Frederiksen, NDRC Project N-106, Memorandum 3, CEEB, May 27, 1943. APP-323-M1 7. Preparation of Norms for the Officer Qualification Test, Form 1, Norman Frederiksen, NDRC Project N-106, Memorandum 4, CEEB, May 28, 1943. APP-319-M3 8. Averages, Standard Deviations, and Intercorrela- tions of Navy Aptitude Tests, Harold Gulliksen, Herbert S. Conrad, and Norman Frederiksen, OSRD 1536, CEEB, June 7, 1943. APP-322-M2 9. Preparation of the United States Navy General Classification Test, Form 1, and the United States Navy Tests of Reading and Arithmetical Reasoning, Form 1, Norman Frederiksen, NDRC Project N-106, Memorandum 7, CEEB, June 23, 1943. APP-327-M1 RESTRICTED 164 BIBLIOGRAPHY 10. Preparation of Experimental Forms of the United States Navy General Classification Test and the United States Navy Tests of Reading and Arith- metical Reasoning, Norman Frederiksen, NDRC Project N-106, Memorandum 8, CEEB, June 28, 1943. 11. Preparation of Norms for Women for the Officer Qualification Test, Form 1, NDRC Project N-106, Memorandum 9, CEEB, July 24, 1943. APP-319-M4 12. Factor Analysis of the New United States Navy Basic Classification Test Battery, D. A. Peterson, OSRD 3004, CEEB, Sept. 29, 1943. APP-412-M1 13. Item Analysis of Navy Aptitude Tests, Herbert S. Conrad, OSRD 3039, CEEB, Dec. 30, 1943. APP-322-M3 14. Development and Validity of the United States Navy Officer Qualification Test, Norman Frederik- sen and D. A. Peterson, OSRD 3186, CEEB, Jan, 7, 1944. APP-319-M5 15. Selection of Test Items by Correlation with an External Criterion, as Applied to the Mechanical Comprehension Test—OQT-O-2, Harold Gullik- sen, OSRD 3187, CEEB, Jan. 8, 1944. APP-325-M2 16. Statistical Analysis of the Mechanical Knowledge Test, Herbert S. Conrad, OSRD 3246, CEEB, Jan. 28, 1944. APP-325-M3 17. Selection of Items for the US Navy General Classi- fication Tests, Form 2, and the US Navy Tests of Reading and Arithmetical Reasoning, Form 2, G. A. Satter, OSRD 3756, CEEB, June 8, 1944. APP-412-M3 18. A Statistical Evaluation of the United States Navy Officer Qualification Test, Forms 2 and 3, OSRD 3978, CEEB, Aug. 4, 1944. APP-319-M6 19. Characteristics and Uses of Item-Analysis Data, Herbert S. Conrad, OSRD 4034, CEEB, Aug. 19, 1944. APP-412-M4 20. The Preparation of Norms for the Fleet Edition of the General Classification Test, D. A. Peterson, OSRD 4242B, CEEB, Oct. 10, 1944. APP-414-M1 21. The Construction and Validation of an Arithmetical Computation Test, G. A. Satter and Norman Frederiksen, OSRD 4556, CEEB, Jan. 8, 1945. APP-411-M2 22. A Statistical Evaluation of the Basic Classification Test Battery, Form 1, Herbert S. Conrad, OSRD 4636, CEEB, May 14, 1945, pp. 77-84. APP-411-M3 23. A Further Study of the Validity of the Arithmetical Computation Test, Norman Frederiksen, OSRD 5302, CEEB, July 3, 1945. APP-411-M4 24. Achievement Exaonination for Signalman School, M. Duane Bown, Thelburn L. Engle, and Harold P. Bechtoldt, OSRD 5460, CEEB, Aug. 20, 1945. 25. Summary Report on Research and Development of the Navy’s Aptitude Testing Program, Herbert S. Conrad, OSRD 6110, CEEB, Oct. 31, 1945. APP-322-M6 26. “The Relation of Item Difficulty and Inter-item Correlation to Test Variance and Reliability,” Harold Gulliksen, Psychometrika, Vol. 10, 1945, pp. 79-91. 27. Program Outlined for Peacetime Enlisted Classi- fication, Enlisted Classification Bulletin, Septem- ber 1945. CHAPTER 15 1. Naval Aptitude Tests: A. Reliability; B. Scoring Accuracy; C. Scoring Procedures, OSRD 1127, CEEB, Dec. 10, 1942. APP-322-M1 2. Selection of Test Items by Correlation with an External Criterion, as Applied to the Mechanical Comprehension Test—OQT-O-2, Harold Gullik- sen, OSRD 3187, CEEB, Jan. 8, 1944. APP-325-M2 3. The Construction and Validation of an Arithmetical Computation Test, G. A. Satter and Norman Frederiksen, OSRD 4556, CEEB, Jan, 8, 1945. APP-411-M2 4. A Statistical Evaluation of the Basic Classification Test Battery, Form 1, Herbert S. Conrad, OSRD 4636, CEEB, May 14, 1945. APP-411-M3 5. A Further Study of the Validity of the Arithmetical Computation Test, Norman Frederiksen, OSRD 5302, CEEB, July 3, 1945. APP-411-M4 6. “Item Selection by Means of a Maximizing Func- tion,” Paul Horst, Psychometrika, Yol. 1, Decem- ber 1936, pp. 229-244. RESTRICTED OSRD APPOINTEES Applied Psychology Panel Chiefs W. S. Hunter C. W. Bray Technical Aides C. W. Bray J. L. Kennedy Dael Wolfle Members G. K. Bennett C. W. Bray Leonard Carmichael C. H. Graham W. S. Hunter J. M. Stalnaker M. S. Viteles Dael Wolfle RESTRICTED 165 CONTRACT NUMBERS, CONTRACTORS, AND SUBJECTS OF CONTRACTS Contract Number Contractor Subject OEMsr-581 The Trustees of Tufts College Medford, Massachusetts Studies of operator performance and sources of error in the operation of field artillery, tank destroyer, and tank guns, and in all types of antiaircraft equipment. OEMsr-614 National Academy of Sciences Washington, D. C. To establish a Committee to carry on research on selection, train- ing, and related problems of manpower use; to study the design of equipment in terms of human capacities; to advise NDRC as to additional research required in the above fields and to recommend to NDRC contractors for carrying on the work. OEMsr-700 The Trustees of the University of Pennsylvania Philadelphia, Pennsylvania Studies in the selection and training of naval gun crews and personnel for destroyers and destroyer escorts; evaluation of methods of training free (aerial) gunners. OEMsr-705 College Entrance Examination Board Princeton, New Jersey Studies and experimental investigations necessary to develop the Navy’s aptitude and achievement testing program. OEMsr-815 Brown University Providence, Rhode Island Studies and experimental investigations in connection with the selection and training of heightfinder and rangefinder opera- tors and fire controlmen; preparation of operating instructions for Navy gun directors; improvement in the design of fire control equipment. OEMsr-830 The Psychological Corporation New York, New York Studies in the selection of men for communication by voice and by radio code; development of methods of training voice and radio code communication personnel; development of a device for transferring Morse code signals to typescript. OEMsr-834 Brown University Providence, Rhode Island Studies of methods of identifying emotionally unstable men prior to their assignment to military duty; studies of the usefulness of Battle Noise Equipment in selecting and training military personnel and retraining psychiatric casualties. OEMsr-919 Yerkes Laboratories of Pri- mate Biology, Inc. Orange Park, Florida Studies of the selection and training of radar operators and the operation of radar equipment. OEMsr-1136 Princeton University Princeton, New Jersey Studies of the selection and training of night lookouts and of night lookout performance. OEMsr-1171 The Regents of the University of Wisconsin Madison, Wisconsin Studies to develop military requirements for flexible gunnery equipment which should be determined by the characteristics of Army gunners; to assist in the development of valid train- ing methods for flexible gunners; to develop adequate methods for the selection and training of personnel for duty in gun directors. OEMsr-1213 The President and Fellows of Harvard College Cambridge, Massachusetts Studies of methods of testing the relative strength of various interests by determining activity preferences. OEMsr-1298 The Trustees of Pennsylvania State College State College, Pennsylvania Studies of job analysis, qualification, and placement of per- sonnel in the amphibious forces. OEMsr-1340 The Board of Trustees of the Leland Stanford Junior University Stanford University, California Studies of methods and devices to aid in the classification and placement of men in naval jobs. OEMsr-1372 The Trustees of the University of Southern California Los Angeles, California Studies in the selection and training of hatchmen and winch- men specialist teams on AKA and APA vessels. 166 RESTRICTED The projects listed below were transmitted to the Office of the Executive Secretary, OSRD, from the War or Navy Depart- ment through either the War Department Liaison Officer for NDRC or the Office of Research and Inventions (formerly the Coordinator of Research and Development), Navy Department. SERVICE PROJECT NUMBERS Service Project Number Title AC-92 Army Projects Investigation of the Ability of Gunners to Learn the Use of Remote Fire Control Sys- tems of Flexible Gunnery. AC-94 Psychological Factors in the Operation of Flexible Gunnery Equipment. SC-67 Training Program in Voice Communication. SC-70 Selection and Training of Radar Operators. SC-88 Methods of Training Radio Code Operators. SOS-6 Study of Operator Performance on All Types of Antiaircraft Equipment. SOS-7 Development of an Activity Preference and Interest Inventory. SOS-11 Psychological Factors in Operation and Design of Field Artillery, Tank Destroyer, and Tank Sighting Equipment. Selection and Training of Height Finder Operators. N-100 Navy Projects Committee on Applied Psychology in the War and Committee on Service Personnel— Selection and Training. N-104 Determination of Reliability, Objectivity, Validity, and Independence of Medical Tests. N-105 Selection and Training of Naval Gun Crews. N-106 Research and Development of the Navy’s Aptitude Testing Program. N-107 Selection and Training of Radio Code Operators. N-109 Selection and Training of Personnel Using Voice Communication Systems. N-lll Psychological Problems in Operation of Antiaircraft Lead Computing Sights and Directors. N-112 Study and Evaluation of Sighting Methods of Instruction Used in U. S. Naval Free Gunnery Training. N-113 Research on a Personal Inventory and Other Tests for Selection for Special Service Tasks. N-114 Selection and Training of Rangefinder and Radar Operators. N-115 Selection and Training of Night Lookouts. N-116 Selection and Training of Personnel. N-117 Job Analysis, Qualification, and Placement of Personnel in the Amphibious Forces. NR-106 Selection and Training of Personnel Assigned to Destroyers and Destroyer Escorts. NS-146 Selection and Training of Radar Operators. NS-366 Development of Morse Code Actuated Printer. RESTRICTED 167 Appendix RESTRICTED FORM 1 UNITED STATES NAVY OFF I C E R QUALIFICATION TEST Read this first; This test is designed to enable you to provide further evidence regarding your abilities. There is no passing or failing score for the teat and no one will be disqualified because of the test score alone. Your score will simply be regarded as further information about you to be added to your application record, Work as well as you can, wasting no time. No one is expected to answer all of the questions correct ly. As a matter of fact, you may answer a substantial number of Items incorrectly and still secure a score that is at least average. Directions for Taking the Teat First, print your name and the other required information on the separate answer sheet furnished with this booklet. Do this now. After filling in these blanks on the answer shea.t, read the rest of these directions. The test consists of three parts. The total time for all three parts is one hour. You may divide your time as you wish, but the following division is recommended as being to your advantage: Part I 15 minutes Part II 15 minutes Part III 25 minutes This schedule leaves 5 minutes to spend on any of the parts which you have not finished, or to check any answers about which you are not sure. Remember, you have one hour to read the directions carefully and to answer as many questions as you can. Directions and sample questions precede the questions for each part. Indicate all your answers on the separate answer sheet. Give only one answer to each question; double answers are graded as incorrect. Do not spend too much time on any one question or on any one part. If you are uncertain, mark the answer you think is most likely to be correct. If you have a question now or at any time during the test, raise your hand and someone will come to you. Make no marks in this test booklet. Indicate all answers on the separate answer sheet. Do not spend too much time on any one question or any one part. Now read the directions for Part I on the next page and start to work at once. As soon as you finish any one page or any one part, go right on to the next without waiting for further instructions. NOTICE: This test is not to be shown, or the contents revealed, to unauthorized persons in or out of the Navy, or reproduced in whole or in part without written authorization from the Bureau of Naval Personnel of the United States Navy. This test booklet must be surrendered with the answer sheet at the close of the test. DO NOT STOP. GO ON TO PART I. RESTRICTED 171 Suggested timet 15 minutes PART I OPPOSITES Directions Each question in this part consists of a CAPITALIZED word followed by five other words numbered from 1 to 5, Decide which one of the five numbered words is most nearly opposite in meaning to the capitalized word. Indicate your choice by making a heavy black mark between the lines under the corresponding number on the separate answer sheet. The questions below have already been marked correctly on the Sample Answer Blank, Study these questions and be sure that you understand how the answers are indicated. A. HOT 1 plain 2 heavy 3 cold 4 soft 5 clean B. LAUGH 1 giggle 2 amuse 3 snort 4 grieve 5 cry C. LUXURY 1 simplicity 2 delicacy 3 flourish 4 dirt 5 elegance Sample Answer Blank Now begin working on OPPOSITES, Mark each answer In the proper apace on the separate answer sheet, Mark only ONE answer to each question. Remember you are to choose the word most nearly opposite in meaning to the CAPITALIZED word. Do not make any marks in this booklet. Work as fast and as accurately as you can. When you finish Part I, go right on to Part II. DO NOT STOP. GO ON TO THE NEXT PAGE. 172 RESTRICTED OPPOSITES Mark only on separate answer sheet. 1. CREDITABLE 1 accidental 2 unworthy 3 invisible 4 menacing 5 vigilant 2. DISTINCT 1 lnfi.rm 2 indivisible 3 vague 4 complex 5 erroneous 3. COGNIZANT 1 uncalled-for 2 deep-seated 3 inadequate 4 original 5 unaware 4. INVALUABLE 1 priceless 2 expensive 3 mercenary 4 vlolable 5 worthless 5. PERTINENT 1 comparable 2 irrelevant 3 courteous 4 melancholy 5 yielding 6. CARDINAL 1 exotic 2 passionate 3 nonessential 4 fraudulent 5 unavowed 7. ZEALOUS 1 dauntless 2 fanatical 3 humane 4 listless 5 unpatriotic 8. EMIT 1 retreat 2 rebound 3 absorb 4 atone 5 crush 9. FLUX 1 stability 2 piety 3 severity 4 elation 5 penury 10. VERIFICATION 1 deprivation 2 experiment 3 inquiry 4 refutation 5 repute 11. DYNAMIC 1 defensive 2 chaotic 3 fruitless 4 apathetic 5 onerous 12. WARY 1 coy 2 stalwart 3 rash 4 lusty 5 shrill 13. MANDATORY 1 unpretentious 2 derogatory 3 discretionary 4 derisive 5 palatable 14. REQUISITE 1 dispensable 2 imperative 3 perquisite 4 prerequisite 5 unrequited 15. PERMEABLE 1 adhesive 2 impervious 3 indulgent 4 negotiable 5 solvent 16. FUNDAMENTAL 1 voluminous 2 approximate 3 perplexing 4 superficial 5 quiescent 17. PHLEGMATIC 1 ungainly 2 ungracious 3 excitable 4 urbane 5 wholesome 18. RECALCITRANCE 1 cooperation 2 resonance 3 restoration 4 relaxation 5 suavity 19. ARBITRARY 1 absurd 2 incompetent 3 productive 4 reasoned 5 unreasoned 20. DISCREPANT 1 fortunate 2 adept 3 consistent 4 obvious 5 intrinsic 21. APPRECIABLE 1 concise 2 deterrent 3 discordant 4 negligible 5 ungrateful 22. EXTRANEOUS 1 extrinsic 2 feasible 3 genuine 4 integral 5 tangible 23. FACILE 1 arduous 2 cursory 3 infamous 4 purblind 5 factitious 24. DIMINUTION 1 exhibition 2 anticipation 3 augmentation 4 elevation 5 transmutation 25. PROPITIOUS 1 auspicious 2 adverse 3 controversial 4 inexperienced 5 disputatious DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 173 26. ASCERTAIN 1 conjecture 2 demur 3 discern 4 dissuade 5 underestimate 2*7 EXPEDITIOUS 1 dilatory 2 elusive 3 habitual 4 latent 5 scrupulous 28. PROSPECTIVE 1 evasive 2 Impending 3 perspective 4 detected 5 realized 29. OBVIATE 1 necessitate 2 dissimulate 3 vilify 4 diversify 5 domineer 30. ACTUATE 1 defy 2 denounce 3 Inhibit 4 obliterate 5 support 31. EFFETE 1 judicious 2 robust 3 placable 4 paltry 5 wonted 32. SALUTARY 1 civilian 2 detrimental 3 brusque 4 nutritious 5 valedictory 33. AUTOMATIC 1 deliberate 2 sporadic 3 ineffectual 4 Inflammable 5 intimate 34. EXPEDIENT 1 admissible 2 devious 3 leIsurely 4 gradual 5 Inadvisable 35. INTELLIGIBLE 1 simulated 2 inscrutable 3 Indescribable 4 ingenuous 5 Innocuous 36. REDUNDANT 1 exalted 2 slack 3 emollient 4 staid 5 succinct 37. EXACERBATE 1 reverberate 2 assuage 3 temporize 4 effervesce 5 coordinate 38. DUCTILE 1 supine 2 refractory 3 pliant 4 Infallible 5 Indefatigable 39. DESICCATED 1 protracted 2 embellished 3 affluent 4 coarsened 5 saturated 40. PIQUANT 1 precise 2 vapid 3 anomalous 4 sagacious 5 munificent 41. ENERVATE 1 intimidate 2 energize 3 horrify 4 ingratiate 5 satiate 42. IMPROVIDENT 1 grudging 2 frugal 3 opportune 4 reverent 5 utilitarian 43. INCLEMENT 1 contrite 2 conventional 3 Implacable 4 lenient 5 Incessant 44. DESULTORY 1 cursory 2 methodical 3 vibrant 4 arrant 5 boisterous 45. SYNOPTIC 1 indicative 2 ocular 3 discursive 4 resultant 5 undeveloped 46. PERFUNCTORY 1 ardent 2 astute 3 constructive 4 erratic 5 peerless 47. LIMPID 1 inflated 2 sonorous 3 vigorous 4 turbid 5 pellucid 48. PALPABLE 1 adamant 2 noisome 3 blatant 4 defunct 5 incorporeal 49. SEDULOUS 1 Indolent 2 agnostlcal 3 uncouth 4 authoritative 5 irascible 50. INTRANSIGENCY 1 culpability 2 license 3 motivation 4 serenity 5 vacillation DO NOT STOP. GO ON TO PART II. 174 RESTRICTED Suggested tine: 15 minutes PART II MECHANICAL COMPREHENSION Directions This part consists of pictures about which questions are asked. Following each question are three answers,. Decide which one of the answers Is correct and make a heavy black mark between the linos under the correspond' Ing number on the answer sheet. The questions below have already been answered correctly on the Sample Answer Blank, Study them and he sure that you understand the problems and how the answers are Indicated, X At which place should one grasp the handle In order to pull out the nail more easily? (1) A (2) B (3) It makes no difference. Y Which bridge Is stronger? (1) A (2) B (3) They are equally strong. Sample Answer Blank Now begin working on MECHANICAL COMPREHENSION. Mark each answer In the proper space on the separate answer sheet. Mark only ONE answer to each question. Do not make any marks In this booklet. Work as fast and as accurately as you can. When you finish Part II, go right on to Part III, DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 175 MECHANICAL COMPREHENSION Mark only on separate answer sheet. Which bridge is stronger? (1) A (2) B (3) They are equally strong. With which handle can one lift the heavier weight? (1) A (2) B (3) It makes no difference. Which radiator will get hot first when a fire is started in the furnace? (1) A (2) B (3) They will get hot at the same time. Which motor boat is approaching the dock correctly for landing? (1) A (2) B (3) Both are approaching correctly. DO NOT STOP. GO ON TO THE NEXT PAGE. 176 RESTRICTED Which ball will roll to the bottom in the shortest time? (1) A (2) B (3) C In which direction is the object more likely to roll? (1) A (2) B (3) It la no more likely to roll in one direction than in the other. Which peavey is correctly placed for rolling the log to the right? (1) A (2) B (3) Both are correctly placed. In which position will the man get a clearer picture of both the nearest and the farthest face? (1) A (2) B (3) The two faces will be equally clear in either position. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 177 Which boy can throw the ball farther? (1) A (2) B (3) They can throw It equally far. ■Which lens would be better for light- ing a fire? (1) A (2) B (3) They are equally good. In which direction will the belt at the right move? (1) A (2) B (3) Neither; the driver will slip. In which direction will shaft X turn? (1) A (2) B (3) Neither; the belt will slip. DO NOT STOP. GO ON TO THE NEXT PAGE. 178 RESTRICTED At which point la the baseball travel- ing faster? (1) A (2) B (3) It travels at the same speed at both points. Against which window does bhe water press harder? (1) A (2) B (3) It presses equally hard agalnab both. Where will the airplane appear to be bo the diver? (1) A (2) B (3) C The bottom of the vessel of water will appear to be nearer which point? (1) A (2) B (3) C DO NOT STOP, GO ON TO THE NEXT PAGE. RESTRICTED 179 Which light will be brighter? (1) A (2) B (3) They will be equally bright. Which switch should be closed first? (1) A (2) B (3) It makes no difference. In which direction la it aafer to turn the aallboat? (1) A (2) B (3) It la equally aafe to turn it in either direction. In which direction la the breeze more likely to be blowing? (1) A (2) B (3) It la equally likely to be blow' Ing In either direction. DO NOT STOP. GO ON TO THE NEXT PAGE. 180 RESTRICTED The train whistle will sound higher In pitch If a man is in which position? (1) A (2) B (3) It will have the same pitch In either position. At which point would It be more dangerous to light a match? (1) A (2) B (3) It would be equally dangerous at both points. Which ten-pound cylinder will roll farther to the right? (1) A (2) B (3) They will roll equally far. Will the solid cylinder A or the hollow cylinder B roll higher up the Incline at the right? (1) A (2) B (3) Both will roll equally high. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 181 Which magnetized needle la farther from the aquator? (1) A (2) B (3) They are the same distance from the equator. Which cart is more likely to tip for- ward when pulled by the rope? (1) A (2) B (3) They are equally likely to tip. Which picture shows how the ball would be reflected in the curved mirror? (1) A (2) B (3) C Which weight would need to he heavier in order to balance the 200-pound load? (1) A (2) B (3) Equal weights would be needed. DO NOT STOP. GO ON TO THE NEXT PAGE. 182 RESTRICTED The weight will be easier to lift if the hook is attached to which ring? (1) A (2) B (3) It will be equally easy with either ring. Which car is more likely to move? (1) A (2) B (3) Both are likely to move. DO NOT STOP. GO ON TO PART III. RESTRICTED 183 Suggested timet 25 minutes PART III ARITHMETICAL REASONING Directions This part consists of arithmetic problems. Following each problem are five answers, lettered from (a) to (o); one of these answers is correct. Solve each problem and indicate your answer by a heavy black mark in the appropriate space on the answer sheet. Use the back of the separate answer sheet for flgurim The answers to the problems below have already been marked on the Sample Answer Blank, Solve these problems for yourself and be sure that you agree. A* If three pencils cost 5 cents, how many pencils can be bought for 15 cents? (a) 5 (b) 7 (c) 9 (d) 10 (e)15 B* Three-fifths of the 240 uniforms for a company were delivered* How many uniforms were there yet to come? (a) 48 (b) 60 (c) 90 (d) 96 (e) 144 Sample Answer Blank Now begin working on ARITHMETICAL REASONING. Mark each answer in the proper space on the separate answer sheet, Mark only ONE answer to each problem. Do not make any marks in this booklet. Use the back of the separate answer sheet for figuring. Work as fast and as accurately as you can. DO NOT STOP. GO ON TO THE NEXT PAGE. 184 RESTRICTED ARITHMETICAL REASONING Mark only on separate answer sheet. Use back of the answer sheet for figuring. 81, The distance between Boston and Halifax is 450 miles. How many Inches would represent this distance on a map whose scale la 50 miles to 1 inch? (a) 3 (b) 4 (c) 5 (d) 9 (e) 27 82, Synthetic rubber coats 40 cents a pound whereas plantation rubber costs 15 cents a pound. The Hospital Corps uses 4,000 pounds a month in adhesive tape. How much do they save in a month on the rubber alone by using plantation rubber? (a) $600 (b) $1,000 (c) $15,000 (d) $60,-000 (e) $100,000 83. A man averages 90 on three arithmetic tests. His scores on the first and second tests were 85 and 94, What was his score on the third test? (a) 89 (b) 89l (c) 89§ (d) 9oi (e) 91 2 3 84, Two submarines start from the same base and travel in the same direction. Submarine X travels' 15 miles an hour and Submarine Y travels li times as fast. How far apart are they at the end of 18 hours? (a) 5 ml. (b) 6 ml. (c) 67| ml. (d) 90 mi, (e) 630 ml. 85, A truck moves 8 feet forward with every revolution of the wheel. At 330 revolutions per minute, what is the speed of the truck in feet per second? (a) 5.5 (b) 41.25 (c) 44 (d) 2,475 (e) 2,640 86. A circular graph represents the classification of enlisted Naval personnel. If 225 out of each 1,000 men are apprentice seamen, how many degrees will represent this group? (A circle has 360°,) (a) 75° (b) 81° (c) 135° (d) 160° (e) 225° 87. Boring the 60 oval holes in a gun cooler requires 3 operations per hole. Ten round holes can be punched in a single operation. If round holes were used, how many operations would be saved on a gun cooler? (a) 14 (b) 170 (o) 174 (d) 186 (e) 1,800 88. A car requires 1,3 gallons of gasoline to go 20 miles at 65 miles per hour and 1 gallon to go the same distance at 25 miles per hour. How many gallons of gasoline will be saved on a trip of 50 miles by traveling at the slower speed? (a) 2 (b) 1 (o) if (d) 2l (e) 2\ 4 4 o 4 89, A room 10 feet wide and 60 feet long is divided into six compartments, each 10 feet square. How many partitions are needed? (a) 5 (b) 6 (o) 7 (d) 59 (e) 60 90. A patrol boat travels due south for 12 miles, then due east for 16 miles. How many miles is it from the starting point? (a) 10.5 (b) 14 (c) 17 (d) 18 (e) 20 DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 185 91. Six seamen can paint 4 boats In 2 days. How many days would it take two seamen to paint 16 boats if they all work at the same rate? (a) lo2 (b) 20 (c) 2l| (d) 24 (e) 28 3 3 92, A reservoir is 60 feet long, 40 feet wide, and of uniform depth, Fow many cubic feet of water must be drawn off to lower the surface 9 Inches? (a) 75 (b) 1,600 (c) 1,800 (d) 3,200 (e) 21,600 93. An assembly process of First Aid kits requiring 300 man-hours was stream- lined to require only 25 minutes. If 5,000 kits were previously assembled in 300 man-hours, how many are now assembled in an hour? (a) 6*2 (b) 720 (c) 2,083i (d) 12,000 (e) 60,000 lo 94, Two pulleys of diameter 6 Inches and 10 Inches are connected by a belt. How many revolutions per minute must the smaller pulley make in order that the larger pulley make 189 revolutions per minute? (a) 68.04 (bj 113.4 (c) 315 (d) 525 (e) 1,134 95, A certain ship-building company can build a ship in 120 days. In how many days must a ship be built in order to increase the company’s output by 33£# ? (a) 80 (b) 90 (c) 116 (d) 160 (e) 180 (d) 160 (e) 180 96, The intensity of Illumination from a given light source varies Inversely as the square of the distance from the source. If a certain searchlight has an intensity of 100,000 foot-candles at 10 feet, what will be the intensity Illumination in foot-candles on an airplane 3,000 feet from the source? (a) ,9 (b) 1.11 (c) 3.33 (d) 9 (e) 333,33 97, A map 6 inches long and Inches wide represents 2,500 square miles of country. On the scale to which the map is drawn an inch represents how many miles? (a) 10 (b) 25 (c) 50 (d) 100 (e) 250 98, A cardboard container with a cover measures 8x2x3 Inches, How many square Inches of cardboard will be saved by making the first two measurements equal if the same capacity is retained? (a) 0 (b) 6 (c) 12 (d) 36 (e) 48 99, Two cubes are made of the same material. The smaller cube weighs 2 pounds, and the larger weighs 128 pounds. The surface area of the larger is how many times that of the smaller? (a) 4 (b) 8 (c) 16 (d) 64 (e) 84 100, What is the largest number of blocks each measuring 3x5x7 inches that can te packed Into a box with inside dimensions of 10x14x25 Inches? (a) 18 (b) 30 (c) 32 (d) 33 (e) 34 RECHECK ANY ANSWERS YOU WISH IN THE REMAINING TIME, 186 RESTRICTED FORM 2 NOTICE; This test is not to be shown, or the contents revealed, to unauthorized persons in or out of the Navy, or reproduced in whole or in part without written authorization from the Bur;-”; of N?va! Personnel of the United State* Navy. This test booklet must be surrendered with the answer sheet at the close of the test. UNITED STATES NAVY OFFICER QUALIFICATION TEST NAVPERS 16561 Read this first; This test Is designed to enable you to provide further evidence regarding your abilities. There is no passing or failing score for the test and no one will be disqualified, because of the test score alone. Your score will simply be regarded as further information about you to be added to your application record. Work as well as you can, wasting no time. No one is expected to answer all of the questions correctly. As a matter of fact, you may answer a substantial number of items incorrectly and still secure a score that is at least average. Directions for Taking the Test This is form 2 of the test as indicated in the upper right hand corner of the booklet; so put a 2 in the space after FORM in the upper right hand corner of the separate answer sheet. Print your name and the other required information on the answer sheet. Do this now. Then read the rest of these directions. The test consists of three parts. The total time for all three parts is one hour. You may divide your time as you wish, but the following division is recommended as being to your advantage: Part I Part II Part III 15 minutes 15 minutes 25 minutes This schedule leaves 5 minutes to spend on any of the parts which you have not finished, or to check any answers about which you are not sure. Remember, you have one hour to read the directions carefully and to answer as many questions as you can. Directions and sample questions precede the questions for each part. Indicate all your answers on the separate answer sheet. Give only one answer to each question; double answers are graded as incorrect. Do not spend too much time on any one question or on any one part. If you are uncertain, mark the answer you think is most likely to be correct. If you have a question now or at any time during the test, raise your hand and someone will come to you. Make no marks in this test booklet. Indicate all answers on the separate answer sheet. Do not spend too much time on any one question or any one part. Now read the directions for Part I on the next page and start to work at once. As soon as you finish any one page or any one part, go right on to the next without waiting for further instructions. DO NOT STOP. GO ON TO PART I. RESTRICTED 187 PART I Suggested tine: 15 minutes OPPOSITES Directions Each question in this part consists of a CAPITALIZED word followed by five other words nunhered from 1 to 5. Decide which one of the five numbered words is most nearly opposite in meaning to the capitalized word. Indicate your choice by making a heavy black mark between the lines under the corresponding number on the separate answer sheet. The questions below have already been marked correctly on the Sample Answer Blank. Study these questions and be sure that you understand how the answers are indicated. A. HOT 1 plain 2 heavy 5 cold 4 soft 5 clean B. LAUGH 1 giggle 2 amuse S snort 4 grieve 5 cry C. LUXURY 1 simplicity 2 delicacy S flourish 4 dirt 5 elegance Sample Answer Blank Now begin working on OPPOSITES, llark each answer in the proper space on the separate answer sheet. Hark only ONE answer to each question. Remember you are to choose the word most nearly opposite in meaning to the CAPITALIZED word. Do not make any marks In this booklet. Work as fast and as accurately as you can. When you finish Part I, go right on to Part II. DO NOT STOP. GO OH TO THE NEXT PAGE. 188 RESTRICTED Mark only on separate answer sheet. OPPOSITES 1. AUTHORIZE 1 devaluate 2 prohibit 5 imitate 4 reprimand 5 rebel 2. MOBILE 1 durable 2 impassive 5 fixed 4 subdued 5 tenacious 3. PERTINENT 1 comparable 2 irrelevant 5 courteous 4 melancholy 5 yielding 4. HARASS 1 spurn 2 loiter 3 implore 4 soothe 5 repair 5. LIABLE 1 cautioned 2 averse 5 exempt 4 Indomitable 5 unqualified 6. INVALUABLE 1 priceless 2 expensive 5 mercenary 4 violable 5 worthless 7. COGNIZANT 1 uncalled-for 2 deep-seated 5 Inadequate 4 secret 5 unaware 8. FLUX 1 stability 2 piety 3 severity 4 elation 5 penury 9. VERIFICATION 1 deprivation 2 experiment 5 inquiry 4 refutation 5 repute 10. JEOPARDIZE 1 accommodate 2 cower 3 negotiate 4 ransom 5 shield 11. EMIT 1 retreat 2 rebound 3 absorb 4 atone 5 crush 12. INITIAL 1 absolute 2 anterior 5 extended 4 distant 5 ultimate 15. RETRENCH 1 revive 2 adjust 3 expand 4 display 5 reflect 14. PRETENTIOUS 1 realistic 2 discouraged 5 convincing 4 modest 5 reliable 15. CAUSTIC 1 bland 2 ample 3 jerky 4 eager 5 flexible 16. INTACT 1 solaced 2 elastic 3 impaired 4 permissible 5 profound 17. FUNDAMENTAL 1 voluminous 2 approximate 5 perplexing 4 superficial 5 quiescent 18. DISTEND 1 heed 2 entice 5 sustain 4 vitiate 5 constrict 19. CORROBORATE 1 disorganize 2 reduce 3 disregard 4 contradict 5 withdraw 20. PERMEABLE 1 adhesive 2 impervious 5 Indulgent 4 negotiable 5 solvent 21. VERACIOUS 1 commendatory 2 scandalous 3 untruthful 4 negative 5 insidious 22. EBULLIENT 1 muddy 2 epigrammatic 3 determined 4 tranquil 5 enigmatic 23. FACILE 1 arduous 2 cursory 5 Infamous 4 purblind 5 factitious 24. EXTRANEOUS 1 extrinsic 2 feasible 3 genuine 4 integral 5 tangible 25. HACKNEYED 1 soggy 2 reserved 3 natural 4 softened 5 original DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 189 26. PROPITIOUS 1 auspicious 2 adverse 3 controversial 4 inexperienced 5 disputatious 27. PHLEGMATIC 1 ungainly 2 ungracious 3 excitable 4 urbane 5 wholesome 28. DISSEMINATE 1 attribute 2 exhilarate 3 reassure 4 reinforce 5 suppress 29. MITIGATE 1 vindicate 2 initiate 3 aggravate 4 encourage 5 predominate 30. ACTUATE 1 defy 2 denounce 3 inhibit 4 obliterate 5 support 31. VITUPERATE 1 laud 2 constrain 3 scrutinize 4 persevere 5 purify 32. SALUTARY 1 civilian 2 detrimental 5 brusque 4 nutritious 5 valedictory 33. INCONGRUOUS 1 infinitesimal 2 enlightening 3 portentous 4 complete 5 appropriate 54. EPHEMERAL 1 enduring 2 infernal 5 immanent 4 violent 5 vulnerable 35. DESULTORY 1 apprehensive 2 methodical 5 vibrant 4 arrant 5 boisterous 36. ABET 1 substantiate 2 abhor 3 dissuade 4 err 5 abstain 37. INTRANSIGENT 1 reconcilable 2 inquisitive 3 rampant 4 sensitive 5 culpable 33. INTREPID 1 contemplative 2 staunch 3 shrewd 4 stoical 5 pusillanimous 39. PERFUNCTORY 1 enthusiastic 2 astute 3 constructive 4 erratic 5 peerless 40. DUCTILE 1 supine 2 refractory 3 pliant 4 infallible 5 indefatigable 41. ANOMALOUS 1 life-like 2 warranted 3 fantastic 4 insignificant 5 normal • 42. IMPROVIDENT 1 grudging 2 frugal 3 opportune 4 reverent 5 utilitarian 43. REDUNDANT 1 exalted 2 slack 3 emollient 4 staid 5 succinct 44. SYNOPTIC 1 indicative 2 ocular 3 discursive 4 resultant 5 undeveloped 45. PERSPICACIOUS 1 devout 2 liberal 3 unimpeachable 4 supercilious 5 obtuse 46. SINUOUS 1 sudden 2 loose 3 fragile 4 straight 5 virtuous 47. TRENCHANT 1 ostensible 2 optimistic 3 unrestricted 4 polished 5 insipid 48. CATEGORICAL 1 chimerical 2 nominal 3 antithetical 4 bombastic 5 equivocal 49. INCLEMENT 1 contrite 2 conventional 5 Implacable 4 lenient S Incessant 50. INDIGENOUS 1 extant 2 wealthy 5 alien 4 fragmentary 5 straightforward DO NOT STOP. GO OH TO PART II. 190 RESTRICTED Suggested tine: 15 minutes PART II MECHANICAL COMPREHENSION Directions This part consists of pictures about which questions are asked. Following each question are three answers. Decide which one of the answers is correct and make a heavy black mark between the lines under the corresponding number on the answer sheet. The questions below have already been answered correctly on the Sample Answer Blank. Study them and be sure that you understand the problems and how the answers are indicated. At which place should one grasp the handle in order to pull out the nail more easily? (1) A (2) B (5) It makes no difference. Which bridge is stronger? (1) A (2) B (5) They are equally strong. Sample Answer Blank Now begin working on MECHANICAL COMPREHENSION. Mark each answer in the proper space on the separate answer sheet. Mark only ONE answer to each question. Do not make any marks in this booklet. Work as fast and as accurately as you can. When you finish Part II, go right on to Part III. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 191 Mark only on separate answer sheet MECHANICAL COMPREHENSION In which case will more paint be required to cover the object? (1) A (2) B (3) They will require the same amount. Which camera is in the better position for taking the picture? (1) A (2) B (?) The positions are equally good. Which diving bell would be safer at great depths? (1) A (2) B (3) One would be as safe as the other. Which bridge is more likely to break in the middle? (1) A (2) B (5) One is just as likely to break as the other. DO NOT STOP. GO ON TO THE NEXT PAGE. 192 RESTRICTED Which boy can throw the ball farther? (1) A (2) B (5) They can throw it equally far. Which kind of pump would you use to inflate a bicycle tire? (1) A (2) B (5) Either kind. Which motorboat is approaching the dock correctly for landing? (1) A (2) B (5) Both are approaching correctly. The sound of the airplane will appear to come from which point? (1) A (2) B (5) C DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 193 Which way should the men tip the table in order to get it intq the next room? (1) A (2) B (5) One way is as good as the other. In which direction is the object more likely to roll? (1) A (2) B (5) It is as likely to roll in one direction as in the other. At which point is the water moving more rapidly? (1) A (2) B (S) The speed is the same at both points. Which pulley is revolving faster? (1) A (2) B (5) They are revolving at about the same speed. DO NOT STOP. GO ON TO THE NEXT PAGE. 194 RESTRICTED Which thermometer will indicate the higher tempera- ture? (1) A (2) B (3) They will show the same temperature. On which screen door is the arrangement of hinges better? (1) A (2) B (3) "They are equally good. Which tire should be placed on the wheel that runs the speedometer in order to show higher speeds? (1) A (2-) B (5) It is not possible to tell. Which man has to pull harder to hold the weight without moving it? (1) A (2) B (3) They will have to pull equally hard. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 195 Which car is more likely to move? (1) A (2) B (5) They are equally likely to move. Which jar contains mercury? (1) A (2) B (3) C To which hook should the eyebolt be attached to make it easier for the man to pull the car? (1) A (2) B (5) It makes no difference. At which point was the stone traveling faster along its path? (1) A (2) B (S) The speed was the same at both points DO NOT STOP. GO ON TO THE NEXT PAGE. 196 RESTRICTED With which windlass could a man lift the heavier weight? (1) A (2) B (5) There is no difference. In which direction is the wind blowing? (1) A (2) B (5) There is no wind. Which jar of water will empty in the shorter time? (1) A (2) B (5) They will empty in the same time. If A is to be raised either by pulling himself up or by having B pull him, who would have to pull the rope harder? (1) A (2) B (5) They will have to pull equally hard. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 197 Which light will be brighter? (1) A (2) B (3) They will be equally bright. Which cylinder will roll higher up the incline at the right? (Disregard friction.) (1) A (2) B (5) They will roll equally high. Which cert is more likely to tip forward when pulled tjy the rope? (1) A (2) B (5) They are equally likely to tip. In which direction will shaft X turn, if the belt is moving in the direction shown at Y? (1) A *2) B (5) Neither; the belt will slip. DO NOT STOP. GO ON TO THE NEXT PAGE. 198 RESTRICTED What size should the lens opening be in order-to get a clear picture of the whole group? (1) Large. (2) Medium. (5) Small. Which magnet will pick up more tacks? (1) A (2) B (5) They will each pick up about the same number. DO NOT STOP. GO ON TO PART III. RESTRICTED 199 Suggested time* 25 ad mate s PART III ARITHMETICAL REASONING Directions This part consists of arithmetic problems. Following each problem are five answers, lettered from (a} to (e); one of these answers is correct. Solve each problem and indicate your answer by a heavy black mark in the appropriate space on the answer sheet. Ose the back of the separate answer sheet for figurine. The answers to the problems below have already been marked on the Sample Answer Blank. Solve these problems for yourself and be sure that you agree. A. If three pencils cost 5 cents, how many pencils can be bought for 15 cents? (a) S (b) 7 (c) 9 (d) 10 (o) 15 B. Three-fifths of the 240 uniforms for a company were delivered. How many uniforms were there yet to come? (a) 48 (b) 60 (c) 90 (d) 96 (e) 144 Sample Answer Blank Now begin working on ARITHMETICAL REASONING. Mark each answer In the proper space on the separate answer sheet. Mark only ONE answer to each problem. Do not make any marks In this booklet. Use the back of the separate answer sheet for figuring. Work as fast and as accurately as you can. DO NOT STOP. GO ON TO THE NEXT PAGE. 200 RESTRICTED ARITHMETICAL REASONING Mark only on separate answer sheet. Use back of the answer Sheet for figuring. 81. In the United States Air Force a squadron consists of 12 to 15 planes. A group consists of 2 to 4 squadrons. What is the minimum number of planes that can make up a group? (a) 24 (b) 27 (c) 56 (d) 41 (e) 60 82. If 5056 of the bombers produced in the United States are transferred to the Allies and if the Allies actually receive 90$ of the bombers transferred to them, what percentage of the bombers produced in the United States are actually received by the Allies? (a) 5% (b) 1056 (c) 27% (d) 60% (e) 120* 85. Synthetic rubber costs 40 cents a pound whereas plantation rubber costs 15 cents a pound. The Hospital Corps uses 4,000 pounds a month in adhesive tape. How much do they save in a month on the rubber alone by using plantation rubber? (a) $600 (b) $1,000 (c) $15,000 (d) $60,000 (e) $100,000 84. Two submarines start from the same base and travel in the same direction. Submarine X travels 15 miles an hour and Submarine I travels times as fast. How many miles apart are they at the end of 18 hours? (a) 5 (b) 6 (c) 6?| (d) 90 (e) 650 85. There are 52 equally spaced directional points on a compass. How many degrees are there between two adjacent points? (Circle = 560*.) <«i£' (c)10if (d)11 i° 86. A mixture contains by weight 1 part of cement, 5 parts of sand, and 5 parts of gravel. How many tons of gravel are needed for 27 tons of this mixture? (b) 51 (c) 9 (d) 15 (e) 481 87. A truck moves 8 feet forward with every revolution of the wheel. At 550 revolutions per minute, what is the speed of the truck in feet per second? (a) 5.5 (b) 41.25 (c) 44 (d) 2,475 (e) 2,640 88. A car requires 1.5 gallons of gasoline to go 20 miles at 65 miles per hour and 1 gallon to go the same distance at 25 miles per hour. How many gallons of gasoline will be saved on a trip of 50 miles by traveling at the slower speed? (a)f (b) 1 (c) if (d) (e) 89. If of a crop is harvested the first day, how many more days will it take to harvest 5 the remainder of the crop? (a) | (b) l| (c) l| (d) l| (e) 2-1 90. A patrol boat travels due south for 12 miles, then due east for 16 miles. How many miles is it from the starting point? (a) 10.5 (b) 14 (c) 17 (d) 18 (e) 20 DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 201 91. The average cost of driving a new car for the first 5,000 miles is 7 cents per mile. For the first 5,000 miles the average is 8 cents. What is the average cost per mile for the remaining 2,000 miles? (a) s|/ (b) 6/ (c) 7/ (d) ?j\i (e) 25/ 92. A reservoir Is 60 feet long, 40 feet wide, and of uniform depth. How many cubic feet of water must be drawn off to lower the surface 9 Inches? (a) 75 (b) 1,600 (c) 1,800 (d) 5,200 (e) 21,600 95. An assembly process of First Aid kits requiring 500 man-hours was streamlined to require only 25 minutes. If 5,000 kits were previously assembled in 500 man-hours, how many are now assembled in an hour? (a) 6-j| (b) 720 (c) 2,083i (d) 12,000 (e) 60,000 94. Two pulleys having diameters of 3 inches and 5 inches are connected by a belt. How many revolutions per minute must the larger pulley make in order that the smaller make 240 revolutions per minute? (Circumference = 2tTr.) (a) 16 (b) 120 (c) 144 (d) 400 (e) 480 95. A certain trip requires 40 minutes at 45 miles per hour. How many minutes would the same trip take at 50 miles per hour? (a) 26 | (b) 531 (c) (d) 55 (e) 60 96. A certain ship-building company can build a ship in 120 days. In how many days must a ship be built in order to increase the company's output by % ? (a) 80 (b) 90 (c) 116 (d) 160 (e) 180 97. The daily edition of a newspaper can be printed in 60 minutes by one press and in 90 minutes by another. In how many minutes could the edition be printed if both presses were running? (a) 50 (b) 56 (c) 57 \ (d) 75 (e) 150 98. A map 6 Inches long and 4^-inches wide represents 2,500 square miles of country. On the scale to which the map is drawn an inch represents how many miles? (a) 10 (b) 25 (c) 50 (d) 100 (e) 250 99. The intensity of illumination from a given light source varies inversely as the square of the distance from the source. If a certain searchlight has an intensity of 100,000 foot-candles at 10 feet, what will be the intensity of illumination in foot-candles on an airplane 5,000 feet from the source? (a) .9 (b) 1.11 (c) 5.35 (d) 9 (e) 553.35 100. Two cubes are made of the same material. The smaller cube weighs 2 pounds, and the larger weighs 128 pounds. The surface area of the larger is how many times that of the smaller? (a) 4 (b) 8 (c) 16 (d) 64 (e) 84 RECHECK AN! ANSWERS YOU WISH IN THE REMAINING TIME. 202 RESTRICTED FORM S NOTICE: This test is not to be shown, or the contents revealed, to unauthorized persons in or out of the Navy, or reproduced in whole or in part without written authorization from the Bureau of Naval Personnel of the United States Navy. This teat booklet must be surrendered with the answer sheet at the close of the test. UNITED STATES NAVY OFFICER QUALIFICATION TEST NAVPERS 16563 Read this first: This test is designed to enable you to provide further evidence regarding your abilities. There is no passing or failing score for the test and no one will be disqualified because of the test score alone. Your score will simply be regarded as further information about you to be added to your application record. Work as well as you can, wasting no time. No one is expected to answer all of the questions correctly. As a matter of fact, you may answer a substantial number of items incorrectly and still secure a score that is at least average. Directions for Taking the Test This is form 5 of the test as indicated in the upper right hand corner of the booklet} so put a 5 in the space after FORM in the upper right hand corner of the separate answer sheet. Print your name and. the other required information on the answer sheet. Do this now. Then read the rest of these directions. The test consists of three parts. The total time for all three parts is one hour. You may divide your time as you wish, but the following division is recommended as being to your advantage: Part I Part II Part III 15 minutes 15 minutes 25 minutes This schedule leaves 5 minutes to spend on any of the parts which you have not finished, or to check any answers about which you are not sure. Remember, you have one hour to read the directions carefully and to answer as many questions as you can. Directions and sample questions precede the questions for each pert. Indicate all your answers on the separate answer sheet. Give only one answer to each question; double answers are graded as incorrect. Do not spend too much time on any one question or on any one pert. If you arc uncertain, mark the answer you think is most likely to be correct. If you have a question now or at any time during the test, raise your hand and someone will come to you. Make do narks in this test booklet. Indicate all answers on the separate answer sheet. Do not spend too much time on any one question or any one part. Now read the directions for Part I on the next page and start to work at once. As soon as you finish any one page or any one part, go right on to the next without waiting for further Instructions. DO NOT STOP. GO ON TO PART I. RESTRICTED 203 PART I Suggested time: 15 minutes OPPOSITES Directions Each question in this part consists of a CAPITALIZED word followed by five other words numbered from 1 to 5. Decide which one of the five numbered words is most nearly opposite in meaning to the capitalized word. Indicate your choice by making a heavy black mark between the lines under the corresponding number on the separate answer sheet. The questions below have already been marked correctly on the Sample Answer Blank. Study these questions and be sure that you understand how the answers are indicated. A. HOT 1 plain E heavy 5 cold 4 soft 5 clean 6. LAUGH 1 giggle E amuse 5 snort 4 grieve 5 cry C. LUXURY 1 simplicity 2 delicacy S flourish 4 dirt 5 elegance Sample Answer Blank Now begin working on OPPOSITES. Hark each answer in the proper space on the separate answer sheet. Mark only ONE answer to each question. Remember you are to choose the word most nearly opposite in meaning to the CAPITALIZED word. Do not make any marks in this booklet. Work as fast and as accurately as you can. When you finish Part I, go right on to Part II. DO NOT STOP. GO ON TO THE NEXT PAGE. 204 RESTRICTED Mark only on separate answer sheet. OPPOSITES 1. ECONOMICAL 1 plentiful 2 selfish 5 cheap 4 scarce 5 wasteful 2. DISTINCT 1 infirm 2 Indivisible S vague 4 complex 5 erroneous 5. ZEALOUS 1 dauntless 2 fanatical 5 humane 4 listless 5 unpatriotic 4. FEASIBLE 1 rigid 2 Impracticable S insufficient 4 clandestine 5 inaccessible 5. PROLONG 1 forestall 2 curtail S disdain 4 pause 5 vanish 6. CONCUR 1 violate 2 criticize 5 dissent 4 vex 5 persist 7. INVALIDATE 1 flatter 2 emulate S entrust 4 claim 5 confirm 8. CONTINGENT 1 urgent 2 antiquated 5 essential 4 Independent 5 continual 9. ACTUAL 1 unknown 2 foreign 5 nysterious 4 imaginary 5 perceptible- 10. COPIOUS 1 Internal 2 intended 5 clever 4 scanty 5 explanatory 11. SELECTIVE 1 Indiscriminate 2 careless 5 unbiased 4 deficient 5 undistinguished 12. DILATE 1 muffle 2 refract 5 complicate 4 contract 5 animate j. ACRID 1 astringent 2 mild 5 refined 4 sodden 5 fluffy 14. DZNAHIC 1 defensive 2 chaotic 5 fruitless 4 apathetic 5 onerous 15. INDUBITABLE. 1 trifling v 2 questionable 5 revocable 4 theoretical 5 indefensible 16. SOMNOLENT 1 Imperious 2 alert S critical 4 scurrilous 5 considerate 17. MANDATORY 1 unpretentious 2 derogatory 5 discretionary 4 derisive 5 palatable 18. REQUISITE 1 dispensable 2 imperative 5 perquisite 4 prerequisite 5 unrequited 19. REPROOF 1 affection 2 laxity 5 sympathy 4 praise 5 suspicion 20. WARI 1 coy 2 stalwart S rash 4 lusty 5 shrill 21. RECALCITRANCE 1 cooperation 2 resonance S restoration 4 relaxation 5 suavity 22. DIMINUTION 1 exhibition 2 anticipation 5 augmentation 4 elevation 5 transmutation 25, DISCREPANT 1 fortunate 2 adept 5 consistent 4 obvious 5 intrinsic 24. APPRECIABLE 1 concise 2 deterrent 5 discordant 4 negligible 5 ungrateful 26. ARBITRARY 1 absurd 2 Incompetent 5 productive 4 reasoned 5 unreasoned DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 205 26. IGNOMINY 1 elegance 2 satisfaction 5 honor 4 order 5 notoriety 27. DELETERIOUS 1 humorous 2 beneficial 3 harmonious 4 aristocratic 5 submissive 28. EXPEDITIOUS 1 dilatory 2 elusive 5 habitual 4 latent 5 scrupulous 29. EFFETE 1 judicious 2 robust 3 placable 4 paltry 5 wonted 30. ASCERTAIN 1 conjecture 2 demur 3 discern 4 dissuade 5 underestimate 31. TRUCULENT 1 overbearing 2 amiable 3 short-sighted 4 reticent 5 niggardly 32. ESTIMABLE 1 contemptible 2 unmistakable 3 inexplicable 4 fabulous 5 meager 33. VOLUPTUOUS 1 diffident 2 ascetic 3 unvaried 4 rapacious 5 fastidious 34. PROSPECTIVE 1 evasive 2 impending 3 perspective 4 detected 5 realized 55. INTELLIGIBLE 1 simulated 2 inscrutable 5 indescribable 4 ingenuous 5 innocuous 36. FATUOUS 1 embarrassed 2 groveling 5 astute 4 matchless 5 relentless 37. OBVIATE 1 necessitate 2 dissimulate 3 vilify 4 diversify 5 domineer 58. ABRIDGE 1 amend 2 amplify 3 demolish 4 improve 5 dissociate 39. SALIENT 1 Implausible 2 explicit 3 paradoxical 4 imperceptible 5 sluggish 40. PIQUANT 1 precise 2 vapid 3 anomalous 4 sagacious 5 munificent 41. EXPEDIENT 1 admissible 2 devious 5 leisurely 4 gradual 5 inadvisable 42. LIMPID 1 inflated 2 sontorous 5 vigorous 4 turbid 5 pellucid 45. ATTENUATE 1 materialize 2 resound 5 Intensify 4 impregnate 5 buoy 44. GERMANE 1 inapplicable 2 illogical 3 figurative 4 salubrious 5 Involved 45. GLIB 1 sedate 2 tactful 3 halting 4 sound S rude 46. ENERVATE 1 intimidate 2 energize 3 horrify 4 ingratiate 5 satiate 47. EXACERBATE 1 reverberate 2 assuage 5 temporize 4 effervesce 5 coordinate 48. PALPABLE 1 adamant 2 noisome 3 blatant 4 defunct 5 Incorporeal 49. PENCHANT 1 displeasure 2 antipathy 5 unconcern 4 exuberance 5 ineptitude 50. EFFULGENT 1 tenebrous 2 Incarnate 5 gelatinous 4 adulterated S osseous DO NOT STOP. GO ON TO PART II. 206 RESTRICTED Suggested timet 15 minutes PART II MECHANICAL COMPREHENSION Directions This part consists of pictures about which questions are asked. Following each question are three answers. Decide which one of the answers is correct and make a heavy black mark between the lines under the corresponding number on the answer sheet. The questions below have already been answered correctly on the Sample Answer Blank. Study them and be sure that you understand the problems and how the answers are indicated At which place should one grasp the handle in order to pull out the nail more easily? (1) A (2) B (S) It makes no difference. Which bridge is stronger? (1) A (2) B (S) They are equally strong. Sample Answer Blank Now begin working on MECHANICAL COMPREHENSION. Mark each answer in the proper space on the separate answer sheet. Mark only ONE answer to each question. Do not make any narks in this booklet. Work as fast and as accurately as you can. When you finish Part II, go right on to Part III. DO NOT STOP. GO OH TO THE NEXT PAGE. RESTRICTED 207 MECHANICAL COMPREHENSION Mark only on separate answer sheet. Which scaffold is more likely to break? (1) A (2) B (S) They are equally likely to break. Which chiael Is better for cutting very hard materiel? (1) A (2) B (5) One is as good as the other. TThich man is more likely to fall if he pushes hard on the drill? (1) A (2) B (5) They ere equally likely to fell. Which bridge Is stronger? (1) A (2) B (5) They are equally strong. DO NOT STOP. GO ON TO THE NEXT PAGE. 208 RESTRICTED Which way should the tiller be pushed in order to approach the dock? (1) A (2) B (5) Either way is correct. Whfch weight is heavier? (1) A (2) B (5) It is impossible to say. In which direction will the belt at the right move? (1) A (2) B (5) Neither} the driving belt will slip. In which direction is the object more likely to roll? (1) A (2) B (5) It is as likely to roll in one direction as in the other. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 209 Which peavey is correctly placed for rolling the log to the right? (1) A (2) B (5) Both are correctly placed. The bottom of the vessel of water will appear to be nearer which point? (1) A (2) B (Z) c With which tool would it be easier to drill hard wood? (1) A (2) B (5) There is no difference. With which handle can one pull the heavier weight out of the well? (1) A (2) B (5) It makes no difference. DO NOT STOP. (50 ON TO THE NEXT PAGE. 210 RESTRICTED Against which window does the water press harder? (1) A (2) B (3) It presses equally hard against both. Which weight puts more strain on the posts? (1) A (2) B (5) There is no difference. Which point -will move faster? (1) A (2) B (5) They will move at the same speed. Which position gives the better lighting for getting something out of the eye? (1) A (2) B (5) They are equally good. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 211 Which thermometer will show the higher temperature? (1) A (2) B (5) They will show the same temperature. If lens X brills the sun’s rays to a focus on the top of the table, in what direction should the other lens be moved so that it will do the same? (1) A (2) B (5) It should not be moved at all. The train whistle will sound higher in pitch if a man is in which position? (1) A (2) B (5) It will have the same pitch in either position. In which tube will the liquid rise? (1) A (2) B (5) It will rise in both tubes. DO NOT STOP. GO ON TO THE NEXT PAGE. 212 RESTRICTED The weight will be easier to lift if the hook is attached to which ring? (1) A (2) B (5) It will be equally easy to lift either way. When the log is lifted by the rope, which end will hang down? (1) A (2) B (5) Neither; the log will remain almost horizontal. Which bell will ring more loudly? (1) A (2) B (5) One will ring as loudly as the other. Which object will roll to the bottom more quickly? (1) A (2) B (3) They will reach it at the same time. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 213 At which point on the dam is the water pressure greater? (1) A (2) B (5) It is the same at both points. Which car is more likely to move? (1) A (2) B (5) They are equally likely to move. In which direction is the wind blowing? (1) A (2) B (5) It is not possible to tell. In which direction will shaft X turn? (1) A (2) B (5) Neither; the belt will slip. DO NOT STOP. GO ON TO THE NEXT PAGE. 214 RESTRICTED Will the solid cylinder A or the hollow cylinder £ roll higher up the incline at the right? (Disregard friction.) (1) A (2) B (S) They will roll equally high. In which case is the pressure on the man's head greater? (1) A (2) B (5) The pressure is exactly the same in both cases. DO NOT STOP. 00 ON TO PART III. RESTRICTED 215 Suggested time: 25 minutes PART III ARITHMETICAL REASONING Directions This part consists of arithmetic problems. Following each problem are five answers, lettered from (a) to (e)j one of these answers is correct. Solve each problem and indicate your answer by a heavy black mark in the appropriate space on the answer sheet. Use the back of the separate answer sheet for flgurl] The answers to the problems below have already been marked on the Sample Answer Blank. Solve these problems for yourself and be sure that you agree. A. If three pencils cost 5 cents, how many pencils can be bought for 15 cents? (a) 5 (b) 7 (c) 9 (d) 10 (e) 15 B. Three-fifths of the 240 uniforms for a company were delivered. How many uniforms were there yet to come? (a) 48 (b) 60 (c) 90 (d) 96 (e) 144 Sample Answer Blank Now begin working on ARITHMETICAL REASONING. Mark each answer in the proper space on the separate answer sheet. Mark only ONE answer to each problem. Do not make any marks in this booklet. Use the back of the separate answer sheet for figuring. Work as fast and as accurately as you can. DO NOT STOP. GO ON TO THE NEXT PAGE. 216 RESTRICTED ARITHMETICAL REASONING Mark only on separate answer sheet. Use back of the answer sheet for figuring. 81. If the passenger rate for air transportation is reduced from 5 cents a mile to 5 cents a mile, how many dollars will be saved on a 1,000 mile trip? (a) $20 (b) $50 (c) $40 (d) $80 (e) $200 82. The distance between Boston and Halifax is 450 miles. How many Inches would represent this distance on a map whose scale is 50 miles to inch? (a) 5 (b) 4 (c) 5 (d) 9 (e) 27 85. In estimating the student-capacity of a class room, 18 square feet of floor space are allowed per student. How many students can be accommodated in a room 36 feet long and 27 feet wide? (a) 5*| (b) 24 (c) 45 (d) 54 (e) 17,496 84. A man averages 90 on three arithmetic tests. His scores on the first and second tests were 85 and 94. What was his score on the third test? (a) 89 (b) (c) 89§ (d) (e) 91 85. In general an invading force should be 50% larger than the defending force to achieve minimum superiority. If the Allies attempt an invasion of Nazi territory, how many divisions should the Allies have to achieve minimum superiority over 50 Nazi divisions? (a) 20 (b) 45 (c) 50 (d) 60 (e) 75 86. A circular graph represents the classification of enlisted Naval personnel. If 225 out of each 1,000 men are apprentice seamen, how many degrees will represent this group? (Circle = 560°.) (a) 75° (b) 81* (c) 155* (d) 160° (e) 225° 87. A room 10 feet wide and 60 feet long is divided into six square compartments, each 10 feet on a side. How many partitions are needed? (a) 5 (b) 6 (c) 7 (d) 59 (e) 60 88. The approximate weight (in pounds) of a shell is one-half the cube of the diameter (in Inches) of the gun. What is the approximate weight in pounds of a shell used in a gun 4 inches in diameter?' (a) 2 (b) 6 (c) 8 (d) 16 (e) 52 89. Step A in manufacturing a certain product can be done by one person at the rate of 25 units every hour. Step B can be done by one person at the rate of 45 units every 2 hours. How many workers will be required at step A to keep 10 workers busy at step B? (a) 9 (b) 9— (c) 11 (d) ll| (e) 18 90. A boat has a speed of 12 miles per hour In still water. How many hours would it take to go 12 ailes down stream when the current is 4 miles per hour? (a) I (b) I (c) 1$ (d) 1-J (e) 5 DO NOT STOP. 00 ON TO THE NEXT PAGE. RESTRICTED 217 91. Six seamen can paint 4 boats in 2 days. How many days would it take two seamen to paint 16 boats if they all work at the same rate? (a) 101 (b) 20 (c) 211 (d) 24 (e) 28 92. Two pulleys of diameter 6 inches and 10 inches are connected by a belt. How many revo- lutions per minute must the smaller pulley make in order that the larger pulley make 189 revolutions per minute? (a) 68.04 (b) 113.4 (c) 515 (d) 525 (o') 1,154 93. A patrol boat, maintaining a constant speed, went due east for an hour and then due north for one hour and twenty minutes, if it then steered a straight return to its base, how many hours was the return trip? («)yf MjfW (d) if (e) zf 94. In 1942 the railroads carried more freight than they carried in 1941 but used only ■g as many cars. The average load per car in .1942 was how many times as great as that for 1941? f (b) g (c) | (d) 1 (e) 2 95, Because of the additional power furnished by a new high-octane gasoline, bombers will be able to increase their bomb loads 5 tons for every 10 tons of bombs now carried. At this rate, how many tons of bombs will a Flying Fortress be able to carry if it can now carry 3 tons of bombs? (a) .9 (b) 5.5 (c) 5.9 (d) 6.35 (e) 10 96. If an American dollar is worth $1.10 in Canadian money, $100.00 in Canadian money is worth how much in American money (to the nearest cent)? (a) $90.00 (b) $90.91 (c) $99.90 (d) $101.00 (e) $110.00 97. A bin 12 feet long, 10 feet wide, and 8 feet high is filled with grain. How many feet must the surface be lowered if 560 bushels are to be removed? (.8 bu.= 1 cu. ft.) (a) 2| (b) 5 (c) 5| (d) (e) 4| 98. A bomber can be assembled in 20 days. If the output is to be increased 25*, in how many days must a bomber be assembled? (a) 15 (b) 16 (c) 18 (d) 25 (e) 26-| 99. Two cubes are to be made of the same material and so constructed that the surface of the larger is 56 times that of the smaller. If the smaller weighs 5 pounds, how many pounds does the larger weigh? (a) 12 (b) 18 (c) 108 (d) 216 (e) 648 100, If 5,000 workers produce 75 pounds of drugs per day, what percentage of the total output is produced by one worker? (a) .015* (b) .02* (c) .67* (d) 1.5* (e) 2* RECHECK ANY ANSWERS YOU WISH IN THE REMAINING TIME. 218 RESTRICTED PERSONAL INVENTORY, LONG FORM Format B DO NOT MARK ON THIS BOOKLET In this questionnaire you are to give information which will help others understand you. You are to indicate certain things about your Job preferences, interests, etc In eaoh question you will always have two answers to choose between — the one on the left side of the page, and the one on the right. Choose the answer which fits you best. Even if neither answer fits you very well, you must ohoose the one that fits you better than the other. Mark your choice on the Answer Sheet, and not on the test booklet. The examiner will show you how to fit the Answer Sheet along side the booklet so that the number of the question you are working on is right next to the same number on the Answer Sheet. Simply indicate your answer by fill- ing in the proper space on the answer sheet. If for a given question you ohoose the answer on the left, then fill in the left-hand space, or if you ohoose the answer on the right, fill in the right-hand space on the Answer Sheet. Do this by making a heavy mark which completely fills the space between the two dotted lines. If you change an answer, be sure to erase completely. Remember, you must always ohoose one answer for each ques tion, but never both. Be sure not to skip any questions. Prepared by Brown University Project N-115 Applied Psychology Panel National Defense Research Committee RESTRICTED 219 1 L 1 prefer a Job that challenges ay ability R I prefer a Job I can do without any difficulty 1 2 I like to take a ohance I like to think things through before I do them 2 3 I hr* "got guts" I can be depended on 3 4 I like to have people do things my nay I like to have people figure things out for me 4 5 The few Jobs I've had I've been very much I've worked at all kinds of things 5 6 interested in I haven't spent much time on a farm Most of my life has been spent on a farm 6 7 I usually plan things ahead and then keep I find that things will usually work them- 7 8 plugging away I prefer working my way up in one Job selves out IVhen I change Jobs I like to go to another town 8 9 I got my first Job near my home town Yes 9 10 I left Jobs because I didn't have to work Yes 10 11 School didn't bother me any more than the I left school because I had enough of it 11 12 next guy In school the other boys went their way I was happiest when I was with the gang 12 13 and I went mine In civilian life I would rather be a In civilian life I would rather be a florist 18 14 mechanic I like office work I like heavier work 14 15 I like working by myself I like working with others 15 16 Since leaving school I've averaged less than one Job a year (or I haven't been out of school a year) Since leaving school I have averaged more than one Job a year 16 17 I'd rather work than take it easy Only a good Job is worth keeping 17 18 I prefer interesting work I prefer a good boss 18 19 I could work bettor if I could get more The average amount of sleep is enough for me 19 20 sleep My employers were more than fair in their I think I was a better worker than my 20 21 opinions of me I have been out of a Job for three months or more in the last five years (or since leaving school) Yes employers thought 21 22 I take life easy I tend to worry 22 23 Being bossed around gripes me A fellow needs a pretty firm boss 23 24 People will step all over you if you Most people will give you a break 24 25 aren't careful Most of my friends are luckier than I am I have my share of lucky breaks 25 Page 1 220 RESTRICTED L R 26 I wish 1 didn't have the blues so often I wish I could be more serious about things 26 27 I miss the comforts I used to have I miss my friends 27 28 I am somewhat disappointed in my family I am proud of my family 28 29 I make friends less easily than the average fellow I make friends more easily than the average fellow 29 30 I left school because 1 preferred working to studying Tea 30 31 I seek excitement I avoid excitement 31 32 When I'm excited or upset 1 feel sweaty When I'm excited or upset I don't show it much 32 33 Once I've made a decision I figure "That's that," and forget it Often after I make a decision I wonder if I've done the right thing 33 34 If I go drinking I can get it out of my system in an evening Once I've started drinking I enjoy going on a real "bat" and getting it out of my system 34 35 36 When I get excited I find it hard to talk straight After exertion I feel hungry When I get excited I talk better than I usually do After exertion I feel dizzy 36 36 37 I pay people back I let people get away with things 37 38 I'm very careful to take medicine when- ever I need it Most people aren't as sick as they think they are 38 39 I like most any kind of food I have a poor appetite 59 40 I've got a steady girl or I'm married I haven’t settled down to one girl yet 40 41 I get embarrassed easily I seldom get embarrassed 41 42 43 I think I might like to watch a surgical operation sometime I like to listen to the radio The sight of blood upsets me I prefer a bang-up party 42 43 44 I want to bo different from my dad I admire my dad's (or guardian’s) way of doing things 44 45 I usually need a drink to get along in a social situation Tec 45 46 The boys know better than to trifle with me I am careful not to get in trouble with the boys 46 47 1 have more headaches than the average person Yes 47 48 I like to stay put I've gone "on the bum" 48 49 I feel nauseated more after eating I feel nauseated more after being excited 49 SO I usually like to go to bed Sometimes I dread going to bed 50 Page 2 RESTRICTED 221 51 L When excited I feel weak R When excited I feel stronger 51 52 I sometimes faint for no good reason Yes 52 53 Being up in high places never bothers me I feel uncomfortable about being in high 53 54 1 can’t think straight when I'm mad places I think better when I'm mad 54 55 I’d like a chance to think without being I prefer being with the boys 55 56 disturbed Drinking alone is usually a waste of good liquor Yes 66 57 58 I think I could have done better in school I think I could have done better in school if 67 if I hadn’t changed schools so often I drink now and then I had studied more I’m proud to say I’ve never touched a drop 58 59 I sometimes drink because it helps me to forget Yes 69 60 I can easily handle a pint of hard liquor without getting dead drunk Yes 60 61 I wish I wouldn’t feel so tired I wish I could have a more responsible job 61 62 I wish I had more responsibility I wish I had more self-confidence 62 63 I wish I weren't so nervous I wish I wouldn’t keep putting things off 63 64 I wish people would stop trying to get I wish I weren’t so affected by what people 64 65 my goat I wish I could get myself to take more chances say I wish worrying wouldn't make me sick to my stomach 65 66 I wish I could have more excitement I wish I weren’t bothered by bad dreams 66 67 I wish I weren't so definite in my I wish people wouldn’t talk me into doing 67 68 opinions I wish I weren’t so different from other things I wish I weren’t so hasty 68 69 people I wish I didn't have so many aches and I wish I wouldn’t keep changing my mind 69 70 pains I wish ray feelings weren't so easily hurt I wish I had more time to spend with my 70 71 I would like to be a baseball player friends I would like to be an artist 71 72 I would rather bo an auto racer I would rather be a poet 72 73 I wish I wouldn’t have so many ups and I wish I could make myself talk more 73 74 downs I’ve always been somewhat uneasy Nothing ever bothers me 74 75 Somehow I never could find enough to do My free time always seemed to bo filled 75 in my free time Page 3 222 RESTRICTED 76 L I think being afraid is worse R I think lying is worse 76 77 I think cops usually do as much harm as good Yes 77 78 I sometimes get confused without any apparent reason Yes 78 79 1 am reckless I am calm 79 80 You have to worry more about heart trouble than most people think Yes 80 81 I wish people would mind their own I wish I could make myself talk more 81 82 business more I am more nervous I am more easy going 82 83 I am more cocky I am more quiet 83 84 I wish I could have more responsibility I wish I wouldn't worry so much 84 85 As a kid I sometimes ran away from home I liked to stay around home too much when I 85 86 As a boy I usually went to my friends* was a kid As a boy I usually brought my friends to my 86 87 houses to play In school things came easy to me house to play In school I had to work for what I got 87 88 In school I had a pretty good idea of what I wanted A lot of things taught in school seemed like a waste of time 88 89 As a child I was often punished when I When I was a child I deserved whatever 69 90 didn’t deserve it My family is somewhat disappointed in me punishment I got My family is fairly pleased with me 90 91 As a boy I had my share of cuts and I had my share of dizzy spells as a boy 91 92 bruises As a kid I played with the older boys As a kid I played with the younger boys 92 93 I didn't let anybody in school put any- In school the boys were usually square with 93 94 thing over on me As a child I had nightmares Yes me 94 95 In school I played hookey more than the In school I played hookey less than the 95 96 average fellow I was a sickly child average fellow I was an active child 96 97 I didn't waste my time hanging around I spent quite a bit of time in school 97 98 after school I have never had a head injury activities I have had a head injury 98 99 In school I often got in trouble with In school I was a little afraid of my 99 100 my teachers In school I liked to be by myself teachers In school I preferred to be with a gang 100 Page 4 RESTRICTED 223 101 L 1 left school because I needed the money Yes R 101 102 I stopped school before I got to high school Yes 102 103 1 graduated from high school Yes 103 104 I got ay first steady job (or entered the armed services) within 4 months after leaving school Yes 104 105 1 repeated a grade in school Yes 106 106 I have felt bad more from staying up I have felt bad more from riding cars. 106 107 late at night 1fy heart sometimes speeds up for no reason at all Yes trains, or busses 107 108 I've had my share of backaches, headaches, and stomach aches Yes 108 109 Being in a warm room makes me drowsy Being in a warm room makes me dizzy 109 110 I have never gone to a doctor for head- I have occasionally gone to a doctor for 110 111 aches or dissy spells I have been annoyed more by sore throat headaches or dizzy spells I have been annoyed more by constipation 111 112 I have felt bad more often from colds and loose bowels I have felt bad more often from nausea 112 113 I have sometimes wet the bed (urinated) I have never wet the bed since the age 113 114 since the age of 10 If I were a research doctor, I would rather find a cure for cancer of 10 If I were a research doctor, I would rather find a cure for epilepsy 114 115 I had an unhappy childhood My childhood was happy 115 11C My own parents brought me up Yes 116 117 My family went its own way pretty much My family took part in a lot of things 117 118 My parents are separated Yes around town 118 119 I was a sensitive kid I was a happy-go-lucky kid 119 120 My parents were away from home a good deal when I was a boy Yes 120 121 I've had my share of sickness I've had my share of happiness 121 122 123 Before entering the armed forces I went to a doctor or a hospital for my nerves I have felt bad more from head cold Before entering the armed forces I never went to a doctor or a hospital for my nerves I have felt bad more from dizziness 122 123 Page 5 224 RESTRICTED 124 L I've always had a good idea of what I'm going to do R Military service seems like a good solution to my problems 124 125 I joined the CCC's more than once Yes 125 126 People talk to each other too much about People are frank and open with me 126 127 things that concern me Having people kid me makes me mad Bumping into something makes me mad 127 128 A friend is a person who helps you get A friend is a person who understands you 128 129 ahead Has anyone in your family ever gone to a doctor or hospital for nervousness? Yes 129 130 My family was on relief for a while Yes 150 131 My family was always quarreling My family stuck too close together 131 132 Our family scraps often came after some- Drinking never was the cause of our family 132 133 one had been drinking My father worked for more than 8 years on one job or for one company Yes scraps 135 134 I have been divoroed Yes 134 135 I am married or intend to get married sometime Yes 135 136 Something makes me say and do things 7 1 can get along in most any situation 136 137 don't mean The hours at night seem long Yes 137 138 Being in a small closed-in place never Being in a small closed-ln place makes me 138 139 bothers me Did you ever wake up and find your tongue bitten? Yes feel uneasy 139 140 Have you bitten your nails since you were 15? Yes 140 141 Have you ever had a convulsion? Yes 141 142 Have you ever been arrested for anything other than a traffic violation? Yes 142 143 Have you ever been sent to jail or reform school? Yes 143 144 Have you ever been a patient in a mental hospital? Yes 144 145 Do you think fighting men should get pensions after the war? Yes 145 Page 6 RESTRICTED 225 OFFICERS' Part B PERSONAL INVENTORY Format 4 Do not write or mark on this booklet. You are to indicate your answers on a separately provided Answer Sheet. In each question you will have two answers to choose between—the one on the left side of the page, and the one on the right. Choose the answer which fits you best, and mark your Answer Sheet accordingly; that is, if you choose the answer on the left, fill in the L space on the Answer Sheet, or if you choose the answer on the right, fill in the R space. Always fill in either one or the other, but never both. Even if neither answer fits you very well, you must choose the one that fits you better than the other. 226 RESTRICTED L R 1 I was an only child Yes No 1 2 As a boy I went to camp I never went to a boys’ camp 2 3 I was a Sea Scout or First Class Boy Scout Yes No 3 4 As a boy I knew Morse Code As a boy I did not know Morse Code 4 5 I couldn’t swim when I was 10 I learned to swim before I was 10 5 6 I was a sickly child I was an active child 6 7 I didn’t waste my time hanging around after school I spent quite a bit of time in school activities 7 8 I was a dare-devil when I was a boy I was overly cautious when I was a boy 8 9 I have grown more tolerant as I have grown older I have grown more cautious as I have grown older 9 10 As a child I had nightmares Yes No 10 11 I was happiest when I was with the gang In school the other boys went their way and I went mine 11 12 As boys we were always thinking up new initiations and hazings We didn’t go in for initiations very much 12 13 As a boy I was ringleader of the gang A friend of mine led the gang 13 14 I was a happy-go-lucky kid In our neighborhood games one of us was always I was a sensitive kid 14 15 getting pretty well banged up I was never injured in any game as a boy 15 16 In school I spent less than half of my free time in athletics In school I spent about half of my free time in ath- letics 16 17 I have coached, or helped coach, the other fellows on the team I’ve never coached, or helped coach, a team 17 IS I consider myself a good athlete I consider myself about an average athlete 18 19 I prefer an opponent of equal skill I prefer playing against an opponent much better than myself 19 20 I enjoy the tired feeling that comes after strenuous exercise I don’t believe in exercising too strenuously 20 21 After exertion I feel dizzy After exertion I feel hungry 21 22 I usually eat a good meal I usually eat very little 22 23 I enjoy camping out and roughing it I’ve never cared much for camping 23 24 At an amusement park I prefer the penny arcade At an amusement park I prefer the roller-coaster 24 25 Most problems are like difficult games Problems and games have little similarity 25 26 Work well done is sufficient reward in itself Work well done should be praised 26 27 I prefer a job that challenges my ability I prefer a job I can do without any difficulty 27 28 I wish I ■wouldn’t feel so tired I wish I could have a more responsible job 28 29 If I have an unpleasant job to do, I try to get it over with as quickly as possible If I have an unpleasant job to do, I like to take my time about doing it 29 RESTRICTED 227 30 L I find any job tiring after a while R I have endurance to resist fatigue and nervous strain 30 31 In civilian life I would rather be a florist In civilian life I would rather be a mechanic 31 32 The sight of blood upsets me I think I might like to watch a surgical operation sometime 32 33 I prefer to be a passenger in an automobile I prefer to be the driver of an automobile 33 34 I have many practical skills I’m better at handling ideas 34 35 I am very much interested in a few things I am interested in all kinds of things 35 36 I prefer working under definite orders I like to be on my own 36 37 I’ve never been particularly mechanically inclined I've always had a mechanical bent 37 38 When things go wrong I call in the repair man I am an expert on fixing things 38 39 I don’t seem to be very inventive I have a couple of schemes that I’ve always felt might make good inventions 39 40 Some of the boys are much better at the job than I am I can do my work as well as anybody 40 41 The average amount of sleep is enough for me I could work better if I could get more sleep 41 42 I never stayed away from work unless I was sick I often had other reasons for staying away from work 42 43 I wish I had more self-confidence I wish I had more responsibility 43 44 I wish I wouldn’t keep putting things off I wish I weren’t so nervous 44 45 I don’t like to attempt a job unless I’m pretty sure of the outcome I’d tackle anything within reason 45 46 When my plans fall through, I usually have another “trick up my sleeve” When my plans fall through, I am usually at a loss as to what to do next 46 47 I make it a point to know my stuff thoroughly I make it a point to know the essentials, but to save a good part of my energy for emergencies 47 48 My ideas are usually unique My ideas are usually practical 48 49 I’m inclined to look at the practical side of things I’m inclined to look at the humorous side of things 49 50 I work best under pressure I work worst under pressure 50 51 I like working with others I like working by myself 51 52 I wish I wouldn’t worry so much I wish I could have more responsibilty 52 53 I never doubt my ability to get things done I never doubt my ability to do things myself 53 54 I prefer to do individual work I prefer to control or handle people 54 55 I find it easier to explain things orally I find it easier to explain things in writing 56 56 I get disgusted when people don’t catch on I get annoyed when people don’t catch on 56 57 I usually ask people to do things I usually tell people to do things 67 58 I am more interested in things I am more interested in people 58 228 RESTRICTED L R 50 I’m more interested in what people are thinking I’m more interested in what people are doing 59 60 People are frank and open with me People talk to each other too much about things that concern me 60 61 Our telephone operators at home often gave me the wrong numbers I never had any troubl e giving information over the telephone 61 62 I prefer to talk to a stranger by phone I prefer to talk to a stranger in person 62 63 I often keep people waiting. I’m always on time for a ppointments 63 64 My free time always seemed to be filled Somehow I never could find enough to do in my free time 64 65 I pick my friends carefully I like to meet new people 65 66 I haven’t settled down to one girl yet I’ve got a steady girl or I’m married 66 67 Most people are o. k. when you get to know them There is always somebody in every crowd that no- body can get along with 67 68 I let people solve their own problems I’m often asked to solve my friends’ problems 68 69 I often make and break little promises A promise should never be broken under any circum- stances 69 70 I remember people’s names easily Ordinarily I don't pay much attention to people’s names 70 71 The second time I meet a person I call him by his first name I don’t call people by their first names unless I know them well 71 72 I would rather have athletic friends I would rather have less active friends 72 73 I write letters mostly to friends I write letters mostly to relatives 73 74 I’m always one of the group Sometimes I feel apart or aloof from the others 74 75 I’m too hard to please I’m too easy to please 76 76 I tend to worry I take life easy 76 77 I avoid excitement I seek excitement 77 78 When excited I feel stronger When excited I feel weak 78 79 I feel nervous more after excitement is over I feel nervous more during stress and excitement 70 80 I can act calm and undisturbed when I am afraid I can’t appear very self-assured when things are un- certain 80 81 I’m looking forward to a crisis to show my stuff I’m not sure just how I would act in a crisis 81 82 I never let myself get in a tight spot When in a tight spot I can always find a way out 82 83 I think better in an emergency I think better when I have time to reflect 83 84 I rarely regret my speedy decisions I regret many of my speedy decisions 84 85 I prefer to make my own decisions I usually seek reliable advice before making decisions 85 86 I’ve always been a deep thinker I’ve always had good judgment 86 87 I can make important decisions immediately I prefer to think things over before deciding 87 88 I usually keep my ideas to myself I like to discuss my ideas with others 88 RESTRICTED 229 89 A good argument provokes thought I dislike any argument 89 90 I prefer to discuss theoretical matters I prefer to discuss practical things 90 91 92 To avoid friction, I accept decisions with which I don’t agree When a party gets dull I liven it up I try to change people around to my point of view I prefer to let others start the ball rolling 91 92 93 I have never led any group I am respected among my friends as a leader 93 94 I can explain things clearly I have a forceful manner 94 95 I’m a hustler I’m a diplomat 95 96 I’m seldom phased Sometimes I get mixed up 96 97 I’m somewhat slow and deliberate in my actions I’m energetic in my actions 97 98 I am more nervous I am more easy going 98 99 I usually show my emotions I usually hide my emotions 99 100 I am calm I am reckless 100 101 Sometimes I dread going to bed I usually like to go to bed 101 102 Our neighbors at home were too noisy Our neighbors at home were too sedate 102 103 When I once get mad, it takes me a while to get over it I flare up easily, but get over it quickly 103 104 I sometimes get confused without any ap- parent reason No Yes 104 105 I think better when I’m mad I can’t think straight when I'm mad 105 106 When I get excited I talk better than I usually do When I get excited I find it hard to talk straight 106 107 I do a certain amount of good wholesome griping I tend to keep my grievances to myself 107 108 I think more quickly than the average person I think about as quickly as the average person 108 109 I enjoy finding my way around in a strange city I prefer being in a city where I know my way around 109 110 I wish worrying wouldn’t make me sick to my stomach I wish I could get myself to take more chances 110 111 I wish I had more time to spend with my friends I wish my feelings weren’t so easily hurt 111 112 People consider me a clear thinker People consider me a good mixer 112 113 People consider me ambitious People consider me unselfish 113 114 I can get along in most any situation Something makes me say and do things I don’t mean 114 115 I tend to be aloof from subordinates I am probably too familiar with subordinates 115 116 I have never tried to teach anybody anything I think I could do a good job of teaching 116 117 Most of my friends are luckier than I am I have my share of lucky breaks 117 118 I criticize my subordinates too much I praise my subordinates too much 118 L R 230 RESTRICTED L R 119 I am stern I am easy going 119 120 Our neighbors at home were too nosey Our neighbors went their own way pretty much 120 121 I’ve had my share of happiness I’ve had my share of sickness 121 122 I wish I could have more excitement I wish I weren’t bothered by bad dreams 122 123 The hours at night seem long Yes No 123 124 I’ve always been somewhat uneasy Nothing ever bothers me 124 125 I sometimes drink because it helps me to forget No Yes 125 126 When I’m excited or upset I don’t show it much When I’m excited or upset I feel sweaty 126 127 Being in a small closed-in place never bothers me Being in a small closed-in place makes me feel uneasy 127 128 I like most any kind of food I have a poor appetite 128 129 I’ve hardly been sick a day in my life I’ve had the average amount of sickness 129 130 I have felt bad more from dizziness I have felt bad more from head cold 130 My heart sometimes speeds up for no reason at 131 all No Yes 131 You have to worry more about heart trouble 132 than most people think No Yes 132 I’ve had my share of backaches, headaches, 133 and stomach aches Yes No 133 134 I’ve had a venereal disease Yes No 134 135 I sometimes faint for no good reason Yes No 135 136 I have more headaches than the average person No Yes 136 137 I have occasionally gone to a doctor for headaches I have never gone to a doctor for headaches or dizzy 137 or dizzy spells spells Someone in my family has gone to a doctor or hos- 138 pital for nervousness Yes No 138 139 I'm very careful to take medicine whenever I need it Most people aren’t as sick as they think they are 139 140 I’d rather be healthy I’d rather be well liked 140 141 I’m happier in the morning I'm happier in the afternoon 141 RESTRICTED 231 142 L Noise and confusion bother me more R Continued silence bothers me more 142 143 I feel uncomfortable about being in high places Being up in high places never bothers me 143 144 I'm more afraid of getting shell shocked I'm more afraid of being physically maimed 144 145 I am more afraid of drowning I am more afraid of being shot 145 146 I like to have things repeated I always hear correctly the first time 146 147 I prefer to r eceive oral orders from my superior I pre fer to receive written orders 147 148 I feel that my education and talents are being wasted right now I feel that the war is making full use of my talents and education 148 149 The Navy is more deserving of praise than the Mer- chant Marine Yes No 149 150 Most of the things I learn in the service are only going to help win the war Most of the things I learn in the service will be of value to me when I get out 150 151 I prefer to have good men under me I prefer to have good officers over me 151 152 Leadership is inborn Leadership can be learned 152 153 154 The chief function of an officer is to keep his men working It is more important for an officer to know the good points of his men The chief function of an officer is to guide his men It is more important for an officer to know the weak points of his men 153 154 155 It is more important for an officer to feel at ease with other officers It is more important for an officer to feel at ease with his men 155 156 157 Officers should welcome advice only from other officers An officer should never be sarcastic Officers should welcome advice from the men in their company Some men will respond only to sarcasm 156 157 158 159 A good leader pitches in Officers should be automatically promoted at sche- duled intervals A good leader arranges things so that it isn't neces- sary for him to pitch in Promotion should come only as a reward for merit 158 159 160 In some ways a scapegoat is an asset in a group Blame can always be correctly placed where it be- longs A sense of humor can aid discipline 160 161 A sense of humor relaxes discipline 161 162 Criticism helps a person do better Criticism makes a person resentful 162 163 Good morale depends more on good men Good morale depends more on good equipment 163 164 Some outfits are more important than others in winning the war Each outfit is equally important in winning the war 164 232 RESTRICTED UNITED STATES NAVY GENERAL CLASSIFICATION TEST FORM I NAVPERS 16500 1943 General Directions Do not write or make marks of any kind in this test booklet. Your answers to the questions in this test are to be marked on the sep- arate answer sheet. Use only the special pencil which has been given you for marking your answers• You are to mark your answer to each question by blacking one of the spaces opposite the number of the question. In black- ing in the space, make several strokes up and down until you have made a glossy black mark. You may erase a mark if you wish to change an answer, but be sure to erase thoroughly. Remember, make no marks on this test booklet. Before you start each part of the test, you will have a chance to read special directions for that part and to work some practice problems. Give only one answer to each question; double answers are counted incorrect. Do not spend too much time on any one question. If you are uncertain, make the best choice you can. Do not skip any questions. Work as rapidly and as accurately as you can. NOTICE: This test is not to be shown, or the contents revealed, to unauthorized persons in or out of the Navy, or reproduced in whole or in part without written authorization from the Bureau of Naval Personnel of the United States Navy. This test booklet must be surrendered with the answer sheet at the close of the test. DO NOT OPEN THIS BOOKLET UNTIL YOU ARE TOLD TO DO SO. RESTRICTED 233 PART I SENTENCE COMPLETION Directions and Samples In this test you will be given sentences to read. One word has been left out of each. Following each sentence are printed five words, one of which Is the word which was left out. You are to decide which one of the five words belongs In the sentence. Look at sample sentence A below: A. The color of milk Is ....... . 1 green 2 red 3 white 4 blue 5 orange The word which was left out of the sentence is white, which is number 3. Therefore the space under 3 has been blacked in bri the separate answer sheet opposite A, in the section labeled “COMPLETION — Samples”, Here is another sample sentence. Find the word which fits in the sentence and black in the proper space on the answer sheet opposite B, Do this now. B. We looked at the clock to see what ••••••• It was. 1 day 2 year 3 weather 4 town 5 time The word which fits in best is number 5, time; so you should have blacked in the space under 5 in line B on the answer sheet. Here are three more sample sentences. Answer them, and mark your answers on the answer sheet. C, I thought he was asleep because his eyes were 1 dark 2 shut 3 dull 4 gray 5 heavy D. A good sailor will the orders of his superior officers. 1 see 2 fear 3 read 4 obey 5 like E. He awoke early in the morning and saw the sinking below the horizon in the west. 1 atm 2 moon 3 storm 4 waves 5 clouds Now check your answers as I give the correct ones to you. ARE THERE ANY QUESTIONS? If you made any mistakes, correct them by thoroughly erasing the wrong answers and marking the answer sheet correctly. On the following pages you will find more sentences like the ones you have Just answered; but some will be more difficult. You should always choose the one word which best fits in with all parts of the sentence. Mark only one answer to each question. If you are not sure of the answer, make the best choice you can. Do not skip any questions. Work as fast and as accurately as you can. STOP HERE. DO NOT GO TO THE NEXT PAGE UNTIL THE SIGNAL "READY BEGIN" IS GIVEN. 234 RESTRICTED SENTENCE COMPLETION Test (Questions Mark only on separate answer sheet. 1. A boat floats on the 1 water 2 sail 3 oar 4 wharf 5 wind 2. A seaman should ....... his duties. 1 decide 2 explain 3 find 4 know 5 divide 3. Always the salute of those under you. 1 approve 2 seek 3 appreciate 4 watch 5 return 4. Most landlubbers are during their first ocean voyage, 1 seasick 2 frightened 3 dissatisfied 4 drowsy 5 amazed 5. After a six months1 voyage, the ship sailed safely Into the ••••.•• • 1 wharf 2 sea 3 port 4 foam 5 sunset 6. The ship is still missing though we have tried for months to it. 1 see 2 restore 3 search 4 locate 5 save 7. When the wind sprang up, the men in the small boat hoisted the • 1 flag 2 sail 3 oars 4 mast 5 signal 8. The men displayed great in the face of heavy gun fire, 1 ability 2 satisfaction 3 agility 4 self-control 5 alacrity 9. He still retains his enthusiasm and courage despite the ••••«.. experiences he has had. 1 varied 2 fatal 3 numerous 4 adventurous 5 harrowing 10. Never stow away a coil of rope unless it is perfectly ••••••., since it will deteriorate quickly if allowed to remain damp. 1 slack 2 waxed 3 dry 4 twined 5 intact 11. He found that it did not pay to be ••••... in disciplinary cases, since the men respected only a commanding officer who was strict with them. 1 harsh 2 lenient 3 unjust 4 wrong 5 fair 12. The fog that sprang up was so dense that although it was he could hardly see his hand before his face. 1 twilight 2 dawn 3 noon 4 night 5 overcast 13. The torpedo is a short-range weapon; to use it the destroyer must get relatively close to its 1 target 2 base 3 fleet 4 explosion 5 tender 14. It was clear in 1942 that victory over Japan would be an victory indeed if it were coupled with a United Nations defeat in Europe at the hands of Germany. 1 Important 2 appalling 3 empty 4 officious 5 indirect 15. Though particular branches of science have appeared to be completely investigated and definitely formulated, this has always proved to be an illusion, , 1 once 2 occasionally 3 never 4 always 5 meanwhile 16. The motor torpedo boats are not even big enough to have .......; the one I went out in was known merely as No. 18. 1 docks 2 turbines 3 reinforcements 4 lifeboats 5 names DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 235 17. Any material placed In the boat must be allowed for by ••••••• the number of passengers. 1 ascertaining 2 Increasing 3 reducing 4 accommodating 5 predicting 10. We patrolled the Channel, skirting the edges of the mine fields and all the time listening for submarines on the ....... device. 1 signalling 2 control 3 aerial 4 detecting 5 depth 19. A vessel navigating In a fog must go no faster than will permit her to within the distance she can see ahead. 1 start 2 stop 3 reverse 4 sail 5 hear 20. The more closely the city and the country are knit together the better; their Interests In the last analysis are 1 practical 2 distinct 3 unselfish 4 necessary 5 mutual 21. It was noticeable that elderly men predominated among the crews; some of the recent recruits, however, were 1 inexperienced 2 relaxed 3 active 4 boys 5 enthusiasts 22. A strategist who adheres Inflexibly to any sot of preconceived rules la hardly likely to be victor against a ....... opponent. 1 resourceful 2 persevering 3 vigilant 4 practiced 5 niggardly 23. Even If one Insists on regarding the air forces as a separate branch of military power It Is a rather ••«.... outlook which concentrates upon the spectacle of the airplane over Its target, to the exclusion of the long chain of circumstances which are responsible for putting It there. 1 Indefinite 2 normal 3 realistic 4 limited 5 cynical 24. The chief disadvantage of the hydrophone was that It picked up other ....... than that of the submarine's propellers. 1 elements 2 features 3 noises 4 clues 5 details 25. In a pursuit as as is the waging of war. It might be set down as an axiom that no one factor, and certainly no one weapon, can be exclusively decisive. 1 difficult 2 complex 3 confused 4 challenging 5 painstaking 26. In recent battleships, the proportion of devoted to armor has gone above the 40 per cent mark. 1 paint 2 Iron 3 tonnage 4 construction 5 space 27. Those to whom the word "defensive" has now become malodorous must be reminded that simultaneous offensives against two powerful and widely removed enemies are usually Impossible and always 1 accidental 2 hazardous 3 overwhelming 4 Interdependent 5 suicidal 28. Despite the great emphasis on protection In the battleship It is armament and not armor that makes a ship. 1 light 2 fast 3 heavy 4 manageable 5 fighting 29. It Is worth noting that armor decks were used on warships long before the development of the airplane, and even now the heaviest Impact which the armor decks are called upon to bear Is from ....... • 1 bombs 2 torpedoes 3 shells 4 depth charges 5 submarines 30. It Is difficult to conceive of gunfire at a target thirteen miles away being more than aerial bombing from only 2,000 feet elevation, but this Is actually the case. 1 accurate 2 Inaccurate 3 harmless 4 risky 5 visible STOP. WAIT FOR FURTHER INSTRUCTIONS. 236 RESTRICTED PART II OPPOSITES Directions and Samples Look at this sample question: P. HOT 1 flat 2 cold 3 light 4 soft 5 wet You know that the opposite of HOT is cold. Cold is number 2, so number 2 has been blacked in on the answer sheet. Look at your answer sheet, in the section labeled "OPPOSITES — Samples", to see how this has been done. Notice that the space under 2 has been blacked in opposite F. In each question in this test you will see a word printed In CAPITAL LETTERS. This word is followed by five numbered words, only one of which means the opposite of the word in capital letters. In each case, you are to find the one word which means the opposite of the word in capital letters. Here is another practice question. Find the answer, and black in the proper space in line G on the answer sheet. G. BUSY 1 tired 2 friendly 3 stupid 4 idle 5 weak The word which means the opposite of BUSY is idle. Idle is number 4, so in line G on the answer sheet, you should have blaclced In the space under number Here are three more practice questions. Do them, and mark your answers on the answer sheet. H. SELFISH 1 generous 2 lazy 3 playful 4 happy 5 beautiful I. BREAK 1 weep 2 find 3 handle 4 use 5 mend J. GRIEF 1 anger 2 poverty 3 Joy 4 sorrow 5 pride Now check your answers while I give the correct ones to you. ARE THERE ANY QUESTIONS? If you made any mistakes, erase your wrong answers completely, and then mark your sheet correctly. In the test on the following pages you will have more questions like the ones you have Just done. Answer each question carefully, but if you are not sure, make the best guess that you can. If you should finish before time is called, go back and chock your answers. DO NOT TURN THIS PAGE UNTIL THE SIGNAL "READY BEGIN" IS GIVEN. RESTRICTED 237 OPPOSITES Test Questions Mark only on separate answer sheet. 31. QUICK 1 strong 2 slow 3 dry 4 dark 5 neat 32. STRAIGHT 1 shiny 2 low 3 crooked 4 clean 5 long 33. CHEERFUL 1 clumsy 2 wise 3 active 4 plain 5 gloomy 34. GIVE 1 lose 2 keep 3 hear 4 live 5 hate 35. SHARP 1 hollow 2 blunt 3 calm 4 heavy 5 clear 36. TRUST 1 blame 2 deny 3 forget 4 doubt 5 surprise 37. REAL 1 foreign 2 imaginary 3 empty 4 faint 5 rare 38. JOIN 1 mislay 2 confuse 3 refute 4 distrust 5 separate 39. CONCEITED 1 afraid 2 rough 3 modest 4 quiet 5 ashamed 40. ADEQUATE 1 improper 2 deceitful 3 insufficient 4 unprotected 5 unnecessary 41. DISCLOSE 1 conceal 2 befriend 3 rectify 4 observe 5 suspect 42. COMPETENT 1 dissatisfied 2 corrupt 3 unfriendly 4 incapable 5 unopposed 43. COMPLIANCE 1 exposure 2 resistance 3 inference 4 reprisal 5 perseverance 44. POLLUTE 1 determine 2 sustain 3 convert 4 dilute 5 purify 45. CONCENTRATE 1 disapprove 2 cheapen 3 diffuse 4 conflict 5 refuse 46. RIGOROUS 1 delinquent 2 superstitious 3 voluntary 4 refined 5 lenient DO NOT STOP. GO ON TO THE NEXT PAGE. 238 RESTRICTED 47. OCCASIONAL 1 frequent 2 former 3 manifold 4 concurrent 5 eventual 48. ACCELERATE 1 punish 2 grovel 3 release 4 soothe 5 retard 49. INVOLVE 1 allay 2 derange 3 suppress 4 exclude 5 dispel 50. IMPAIR 1 enhance 2 damage 3 complement 4 sanction 5 acquiesce 51. INDISPENSABLE 1 compensatory 2 superfluous 3 exceptional 4 reprehensible 5 conditional 52. METICULOUS 1 slovenly 2 prudish 3 indiscreet 4 confiding 5 gluttonous 53. CONCERTED 1 unmusical 2 discouraged 3 inapplicable 4 disorganized 5 mistaken 54. VINDICTIVE 1 irascible 2 unselfish 3 forgiving 4 reactive 5 communicative 55. LATENT 1 stunted 2 beneficent 3 secular 4 apparent 5 mismanaged 56. BLATANT 1 astute 2 deliberate 3 jubilant 4 reticent 5 melancholy 57. DISPARATE 1 reciprocal 2 moderate 3 appropriate 4 concomitant 5 identical 58. EQUABLE 1 unjustified 2 bewildering 3 unstable 4 eccentric 5 callous 59. INIMITABLE 1 fabulous 2 ordinary 3 Jocund 4 loquacious 5 mitigated 60. RECANT 1 avow 2 congeal 3 covet 4 expunge 5 elicit STOP. WAIT FOR FURTHER INSTRUCTIONS, RESTRICTED 239 PART III ANALOGIES Directions and Samples In this test you will be asked to find certain relationships between words. You will have a number of sentences to read; in each sentence you will have one pair of words, such as NIGHT and dark, which have the same relationship as another pair of words, such as DAY and light. Read the following sentence: During the NIGHT it is dark, but during the DAY it is light« Notice that "night*1 is related to "dark" in the same way that "aay" is related to "light", so we may say that NIGHT is to dark as DAY is to light. Now look at this one: You READ a book and LISTEN to music. Here you can see that" READ is to book as LISTEN is to music. In the question labeled K below, you are to select the one of the five numbered words which besT completes the thought. iC. READ la to book as LISTEN is to 1 music 2 air 3 pencil 4 newspaper 5 platform The correct answer is "music", or number 1, Notice on the answer sheet, in the section labeled "ANALOGIES — Samples", that space 1 has been blacked in, opposite the letter K. What is the correct answer to the following question? L. GLOVE is to hand as HAT is to 1 face 2 fingers 3 forehead 4 body 5 head A GLOVE is worn on the hand, and a HAT is worn on the head; so head, or number 5, is the correct answer. Show the answer by blacking In the space under 5, opposite the letter L, on the answer sheet. Do this now. Find the answers to the following questions yourself; record your answers on the answer sheet. M. PISH is to awlm as BIRD Is to 1 air 2 fly 3 feathers 4 ride 5 wings N. WATER is to sponge as INK is to 1 pen 2 'bottle 3 write 4 blotter 5 desk 0. THERMOMETER is to temperature as CLOCK is to 1 month 2 heat 3 time 4 hand 5 dial DO NOT STOP. GO ON TO THE NEXT PAGE. 240 RESTRICTED Now check your answers as I give the correct ones to you. ARE THERE ANY QUESTIONS? If you made any mistakes, erase your wrong answers completely, and then mark your sheet correctly. In the test on the following pages you will have more questions like the ones you have just done. Answer each question carefully, but if you are not sure, make the best guess that you can. If you should finish before time Is called, go back and check your answers. DO NOT TURN THIS PAGE UNTIL THE SIGNAL "READY BEGIN" IS GIVEN. RESTRICTED 241 ANALOGIES Mark only on separate answer sheet. 61. EYE is to see as EAR is to 1 nose 2 head 3 hoar 4 speak 5 listen 62. TREE is to branch as HAND is to 1 arm 2 glove 3 wrist 4 finger 5 fist 63. DOOR is to room as LIPS are to 1 tongue 2 mouth 3 nose 4 teeth 5 face 64. GUN is to atm as BOAT is to 1 rudder 2 float 3 steer 4 sail 5 navigate 65. FISH is to scales as BIRD is to 1 wings 2 feet 3 eggs 4 feathers 5 skin 66. HOUSE is to foundation as TREE is to 1 garden 2 'branches 3 roots 4 bark 5 leaves 67. SHIP is to lifeboat as PLANE is to 1 safety halt 2 hangar 3 aircraft carrier 4 glider 5 parachute 68. FOREST is to trees as NAVY is to 1 sailors 2 admirals 3 defense 4 food 5 ocean 69. CROP is to land as FISH is to 1 fisherman 2 sea 3 market 4 fish-net 5 boat 70. AIR is to wind as WATER is to 1 land 2 waves 3 clouds 4 ocean 5 fish 71. ASK is to receive as ATTEMPT is to 1 desire 2 give 3 begin 4 wait 5 succeed 72. TRAINS are to passengers as BOOKS are to 1 book-stores 2 authors 3 paper 4 publishers 5 ideas 73. THROY/ is to catch as TEACH is to 1 remember 2”study 3 learn 4 read 5 respect 74. GROUND is to worm as OCEAN is to 1 submarine" 2 destroyer 3 cable 4 sea 5 depth bomb 75. LAVA is to volcano as SPEECH is to 1 words 2 orat'or 3 audience 4 eloquence 5 temper 76. SALARY is to service as PENALTY is to 1 work 2 disapproval 3 discipline 4 misdemeanor 5 pain 77. MATCH is to flame as DETONATOR is to 1 explosion 2 trigger 3 bullet 4 gun 5 powder 78. WATER is to island as LAND is to 1 continent 2"penlnsula 3 ocean 4 lake 5 stream 79. TROUBLE is to unhappiness as SUN is to 1 heavens 2 earrVi !5 warmth 4 cold 5 moon DO NOT STOP. GO ON TO THE NEXT PAGE. 242 RESTRICTED 80. LITERATURE is to book as ART is to 1 artist 2 painting 3 color 4 beauty 5 scenery 81. INVENTION is to idea as RIVER is to 1 ocean 2 flood 3 crops 4 spring 5 bank 82. POSITIVE is to uncertain as KNOWLEDGE is to 1 logic 2 faltH STproof 4 superstition 5 convention 83. CREST is to trough as HILL is to 1 mountain 2 meadow 3 slope 4 lake 5 valley 84. AIRPLANE is to wings as AUTOMOBILE is to 1 motor 2 land 3 wheels 4 brake 5 hood 85. INTERMITTENT is to continuous as FLICKER is to 1 frighten 2 blink 3 fade 4 glare 5 glint 86. THREATEN is to fulfill as CLOUDS are to 1 winter 2 darkness 3 storm 4 dampness 5 sunshine 87. TWILIGHT is to darkness as SORROW is to 1 misfortune 2 despair 3 discomfort 4 happiness 5 regret 88. MORNING is to wake as SPRING is to 1 flowers 2 thaw 3 grow 4 sprout 5 welcome 89. MONTH is to time as QUART is to 1 gallon 2 measure 3 volume 4 peck 5 scales 90. CHARACTER is to reputation as REALITY is to 1 appearance 2 conduct 3 Integrity 4 wisdom 5 possibility 91. STREAM is to torrent as RETREAT is to 1 rout 2 victory 3 failure 4 army 5 battle 92. TELESCOPE is to lens as VIOLIN is to 1 violinist 2 Vow 3 strings 4 music 5 piano 93. LIGHTHOUSE is to reef as CONSCIENCE is to 1 man 2 conduct 3 good 4 soul 5 temptation 94. CONTINENT is to isthmus as OCEAN is to 1 channel 2 hay 5 peninsula 4 strait 5 lake 95. BLOTTER is to ink as EGOTIST is to 1 friends 2 enemies 3 selfishness 4 contempt 5 praise 96. DIVE is to fall as SWIM la to 1 sink 2 f'loat 3 drown 4 rise 5 rescue 97. MEDICINES are to health as FERTILIZERS are to 1 riches 2 soil 5 agriculture 4 farm 5 productivity 98. MATCH is to fuse as FUSE is to 1 explosion 2 gun 3 bullet 4 powder 5 trigger 99. PEN is to ink as SUBMARINE is to 1 torpedo 2 explosion 3 dive bomber 4 ocean 5 ship 100. SAPLING is to lumber as CALF la to 1 cow 2 bee? 5 shoes 4 bull 5 veal STOP. WAIT FOR FURTHER INSTRUCTIONS. RESTRICTED 243 UNITED STATES NAVY GENERAL CLASSIFICATION TEST FORM 2 NAVPERS - 16502 1943 General Directions Do not write or make marks of any kind in this test booklet. Your answers to the questions in this test are to be marked on the sep- arate answer sheet. Use only the special pencil which has been given you for marking your answers. You are to mark your answer to each question by blacking one of the spaces opposite the number of the question. In black- ing in the space, make several strokes up and down until you have made a glossy black mark. You may erase a mark if you wish to change an answer, but be sure to erase thoroughly. Remember, make no marks on this test booklet. Before you start each part of the test, you will have a chance to read special directions for that part and to work some practice problems. Give only one answer to each question; double answers are counted as incor- rect. Do not spend too much time on any one question. If you are uncertain, make the best choice you can. Do not skip any questions. Work as rapidly and as accurately as you can. NOTICE: This test is not to be shown, or the contents revealed, to unauthorized persons in or out of the Navy, or reproduced In whole or In part without written authorization from the Bureau of Naval Personnel of the United States Navy. This test book- let must be surrendered with the answer sheet at the close of the test. DO NOT OPEN THIS BOOKLET UNTIL YOU ARE TOLD TO DO SO. RESTRICTED 245 PART I SENTENCE COMPLETION Directions and Samples In this test you will be given sentences to read. One word has been left out of each. Following each sentence are printed five words, one of which Is the word which was left out. You are to decide which one of the five words belongs in the sentence. Look at sample sentence A below: A. The color of milk is 1 green 2 red 3 white 4 blue 5 orange The word which was left out of the sentence is white, which is number 3. Therefore the space under 3 has been blacked in on the separate answer sheet opposite A, In the section labeled ’’COMPLETION — Samples". Here is another sanqple sentence. Find the word which fits in the sentence and black in the proper space on the answer sheet opposite B. Do this now. B, We looked at the clock to see what it was. 1 day 2 year 3 weather 4 town 5 time The word which fits in best Is number 5, time; so you should have bracked in the space under 5 In line B on the answer sheet. Here are three more sample sentences. Answer them, and mark your answers on the answer sheet. C. I thought he was asleep because his eyes were 1 dark 2 shut 3 dull 4 gray 5 heavy D. A good sailor will the orders of his superior officers, 1 see 2 fear 3 read 4 obey 5 like E. He awoke early in the morning and saw the sinking below the horizon in the west. 1 sun 2 moon 3 storm 4 waves 5 clouds Now check your answers as the correct ones are given to you. ARE THERE ANY QUESTIONS? If you made any mistakes, correct them by thoroughly erasing the wrong answers and marking the answer sheet correctly. On the following pages you will find more sentences like the ones you have just answered; but some will be more difficult. You should always choose the one word which best fits in with all parts of the sentence. Mark only one answer to each question. If you are not sure of the answer, make the best choice you can. Do not skip any questions. Work as rapidly and as accurately as you can. 2 DO NOT GO ON TO THE NEXT PAGE UNTIL THE SIGNAL "READY - BEGIN" IS GIVEN. 246 RESTRICTED SENTENCE COMPLETION Test Questions Mark only on separate answer sheet. 1. The main task of a navy is to win . 1 ships 2 guns 3 prizes 4 friends 5 battles 2. It takes a lon'g-time to a battleship. 1 build 2 defend 3 float 4 stop 5 attack 3. Strong electric currents can have such a effect on the body that the injury they cause may be permanent. 1 mysterious 2 harmful 3 healthful 4 thrilling 5 mild 4. A navy is large if it has many 1 duties 2 officers 3 oceans 4 ships 5 parts 5. When a steam and a sailing vessel are proceeding so as to involve risk of , the steam vessel shall keep out of the way of the sailing vessel. 1 upset 2 collision 3 sinking 4 grounding 5 stopping 6. The air escaping from the old type of torpedo as it traveled through the water marked its so that it could sometimes be avoided by its intended victim. 1 depth 2 path 3 size 4 source 5 distance 7. The men must learn to fresh water as much as possible, as it is obtained at great expense on board ship, 1 purify 2 drink 3 conserve 4 enjoy 5 secure 8. Because the men had taken their shoes off, their feet became from wading up onto the stony beach. 1 tired 2 wet 3 hot 4 cold 5 bruised 9. A few torpedo hits on Important ships of a retreating fleet may force its admiral into an action he would otherwise 1 seek 2 win 3 avoid 4 lose 5 dislike 10. The .... of a bullet shot from a modern rifle is such that the bullet may strike a man before he hears the sound of the explosion which sent it on its way. 1 Impact 2 path 3 speed 4 effect 5 sound 11. The captain doubted whether the submarine's ballast tanks were still intact; if they were not, there was no hope of blowing the water out of them and thereby losing enough to allow the boat to return to the surface. 1 weight 2 air 3 speed 4 bulk 5 descent 12. Supplies were getting scarce and the torpedo boats were ....... to make further offensive raids since we couldn't afford to risk the boats or spare the gas. 1 forbidden 2 needed 3 forced 4 prepared 5 used 13. In naval warfare, the side which sees that it is clearly weaker will usually attempt to action. 1 begin 2 evade 3 win 4 settle 5 organize 14. Sailing vessels may seem to be very to operate, but due to the cost of handling sails they are more expensive than steam-driven freighters. 1 difficult 2 easy 3 cheap 4 slow 5 costly 15. TKT is a powerful and dangerous explosive, but it can be handled pro- vided its characteristics are known. 1 carefully 2 skillfully 3 cheaply 4 speedily 5 safely DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 247 16. The correct way to remove the insulation from heavy wire is to the insulation off the wire as you would sharpen a lead pencil. 1 peel 2 whittle 3 pull 4 pick 5 rub 17. Although the ship was very , he could see every detail because the air was unusually clear. 1 close 2 vague 3 large 4 distant 5 indistinct 18. The importance in warfare of ability to transport supplies is generally recognized; however, upon the entrance of Japan and the United States into the war the daily press was full of comparisons of the naval strength of the two countries but entirely neglected the matter of tonnage. 1 lost 2 portable 3 shipping 4 heavy 5 utilizing 19. On board ship, the of time is absolutely necessary because it may be impossible to determine a ship's position unless the time of day is known, 1 planning 2 passage 3 saving 4 measurement 5 minimum 20. The ....... of the explosive charge in a shell is not a good indication of its destructive power because some explosives are much more powerful pound for pound than others. 1 structure 2 weight 3 size 4 danger 5 appearance 21. Much of the criticism of the government arose because surface developments were taken as indications of basic policy when they were in fact merely steps in the process of achieving that policy. 1 harmless 2 preliminary 3 general 4 alleged 5 controversial 22. There is a saturation point of supply, and all above that successful operations. 1 determines 2 prolongs 3 hinders 4 paralyzes 5 eases 23. Despite the defenders' resort to the "scorched earth" policy in Malaya, the Japanese vast stores of tin and rubber with possibilities of later exploitation. 1 attacked 2 had 3 destroyed 4 maintained 5 acquired 24. Wire rope of the same size as Manila rope is much stronger, so that for a given task a wire rope may be used which is than the Manila rope which would be needed for the same purpose. 1 thinner 2 tougher 3 thicker 4 stlffer 5 stronger 25. The formation of rust la by heat, as is shown by the fact that the parts of a ship's hull near the boilers rust more rapidly than other parts. 1 damaged 2 accompanied 3 delayed 4 Increased 5 caused 26. In the majority of cases gas poisoning is due to carbon monoxide, and unless it is known definitely that some other poisonous gas is responsible, it is usually that carbon monoxide is the cause of asphyxiation. 1 known 2 clear 3 found 4 supposed 5 doubted 27. Care is sometimes necessary when surprise of the enemy is needed, but at other times caution may actually result in greater than would boldness. 1 losses 2 victories 3 advantages 4 effectiveness 5 bravery 28. The great commander must be able to stick to his course despite a thousand distractions, and yet be stifficiently to recognize when a change in circumstances demands a change in plan. 1 agreeable 2 informed 3 flexible 4 Independent 5 bold 29. Intensified enemy submarine operations in the western Atlantic soon put in- creased strain upon shipping that had already been by a greater military demand for transports. 1 taxed 2 expanded 3 weakened 4 necessitated 5 required 30. The propaganda expert is convinced that the psychological front is decisive in modern warfare, the economist is equally certain that the production effort is most important, and among military men there will be all kinds of divisions of opinion according to the of the disputants. 1 wisdom 2 characteristics 3 ambitions 4 intelligence 5 specialties STOP. WAIT FOR FURTHER INSTRUCTIONS. 248 RESTRICTED PART II OPPOSITES Directions and Samples Look at this sample question: P. HOT 1 flat 2 cold 3 light 4 soft 5 wet You know that the opposite of HOT is cold. Cold is number 2, so number 2 has been blacked in on the answer sheet. Look at your answer sheet, in the section labeled "OPPOSITES — Samples," to see how this has been done. Notice that the space under 2 has been blacked in opposite P. In each question in this test you will see a word printed in CAPITAL LETTERS. This word is followed by five numbered words, only one of which means the opposite of the word in capital letters. In each case, you are to find the one word wnlch means the opposite of the word in capital letters. Here is another practice question. Find the answer, and black in the proper space in line G on the answer sheet. G. BUSY 1 tired 2 friendly 3 stupid 4 idle 5 weak The word which means the opposite of BUSY is idle. Idle is m’mber 4, so in line G on the answer sheet, you should have blacked in the space under number 4. Here are three more practice questions. Do them, and mark your answers on the answer sheet. H. SELFISH 1 generous 2 lazy 3 playful 4 happy 5 beautiful I. BREAK 1 weep 2 find 3 handle 4 use 5 mend J. GRIEF 1 anger 2 poverty 3 Joy 4 sorrow 5 pride Now check your answers as the correct ones are given to you. ARE THERE ANY QUESTIONS? If you made any mistakes, erase your wrong answers completely, and then mark your sheet correctly. In the test on the following pages you will have more questions like the ones you have Just done. Answer each question carefully, but if you are not sure, make the best guess that you can. If you should finish before time is called, go back and check your answers. DO NOT TURN THIS PAGE UNTIL THE SIGNAL "READY - BEGIN" IS GIVEN. RESTRICTED 249 OPPOSITES Test Questions Mark only on separate answer sheet. 31. EARLY 1 past 2 slow 3 lazy 4 late 5 future 32. ANGRY 1 pleased 2 hopeful 3 careful 4 comfortable 5 boastful 33. FIND 1 send 2 leave 3 lose 4 drop 5 hide 34. APPROACH 1 extend 2 dismiss 3 attack 4 withdraw 5 rebel 35. FASTENED 1 abandoned 2 loose 3 wide 4 broken 5 exposed 36. INTERESTING 1 serious 2 unpopular 3 dull 4 difficult 5 foolish 37. MATURE 1 undeveloped 2 partial 3 weak 4 inferior 5 ridiculous 38. PROBABLE 1 possible 2 unsuitable 3 concealed 4 unlikely 5 unfavorable 39. COMPULSORY 1 superfluous 2 voluntary 3 enthusiastic 4 Indulgent 5 desired 40. GRADUAL 1 unexpected 2 irregular 3 greedy 4 excessive 5 sudden 41. YIELD 1 Intend 2 repudiate 3 provoke 4 manage 5 withstand 42. EXCLUDE 1 connect 2 pamper 3 admit 4 repay 5 support 43. BRISK 1 sluggish 2 coarse 3 vague 4 shy 5 smooth 44. FACILITATE 1 hinder 2 obey 3 camouflage 4 coerce 5 clarify DO NOT STOP. GO ON TO THE NEXT PAGE. 250 RESTRICTED 45. SYNTHETIC 1 moist 2 intense 3 crude 4 normal 5 natural 46. INTRICATE 1 neat 2 simple 3 complex 4 vast 5 soiled 47. DEPLETE 1 acquire 2 pump 3 diffuse 4 replenish 5 value 48. PROMINENT 1 unessential 2 pitiful 3 inconspicuous 4 contemptible 5 futile 49. RIGID 1 curved 2 crooked 3 pliant 4 chaotic 5 floating 50. ANONYMOUS 1 identified 2 authentic 3 egotistic 4 meaningful 5 outstanding 51. STRINGENT 1 lax 2 muffled 3 subtle 4 embellished 5 fragrant 52. PRECARIOUS 1 secure 2 bold 3 convincing 4 deep 5 weighty 53. PRECEDE 1 proceed 2 depart 3 oppose 4 follow 5 claim 54. SUBSEQUENT 1 ultimate 2 prior 3 retarded 4 superb 5 derivative 55. CESSATION 1 encroachment 2 variation 3 continuation 4 creation 5 acquisition 56. CIRCUITOUS 1 concrete 2 conclusive 3 brief 4 direct 5 compact 57. UNPRECEDENTED 1 awaited 2 customary 3 premature 4 resultant 5 basic 58. INCENTIVE 1 contradiction 2 proposal 3 requirement 4 expenditure 5 deterrent 59. LUCRATIVE 1 unprofitable 2 parsimonious 3 emollient 4 tedious 5 economical 60. AUGMENT 1 conciliate 2 pollute 3 diminish 4 repress 5 exhaust STOP. WAIT FOR FURTHER INSTRUCTIONS. RESTRICTED 251 PART III ANALOGIES Directions and Samples In this test you will be asked to find certain relationships between words. You will have a number of sentences to read; in each sentence you will have one pair of words, such as NIGHT and dark, which have the same relationship as another pair of words, such as DAY and light. Read the following sentence: During the NIGHT it is dark, but during the DAY it is light Notice that "night'' is related to "dark" in the same way that "day" is related to "light", so we may say that NIGHT is to dark as DAY is to light. Now look at this one: You READ a book and LISTEN to music. Here you can see that READ is to book as LISTEN is to music. In the question labeled K below, you are to select the one of the five numbered words which best completes the thought. K. READ is to book as LISTEN is to 1 music "2 air 3 pencil 4 newspaper 5 platform The correct answer is "music", or number 1. Notice on the answer sheet, in the section labeled "ANALOGIES -- Samples", that space 1 has been blacked In, opposite the letter K. What is the correct answer to the following question? L. GLOVE is to hand as HAT la to 1 face 2 fingers 3 forehead 4 body 5 head A GLOVE is worn on the hand, and a HAT is worn on the head; so head, or number 5, Is the correct answer. Show the answer by blacking in” the space under 5, opposite the letter L, on the answer sheet. Do this now. Find the answers to the following questions yourself; record your answers on the answer sheet. M. PISH is to swim as BIRD is to 1 air 2 fly 3 feathers 4 ride 5 wings N. WATER la to sponge as INK is to 1 pen 2 Bottle 3 write 4 blotter 5 desK 0. THERMOMETER is to temperature as CLOCK is to 1 month 2 heat 3 time 4 hand 5 dial Now check your answers as I give the correct ones to you. ARE THERE ANY QUESTIONS? DO NOT STOP. GO ON TO THE NEXT PAGE. 252 RESTRICTED If.you made any mistakes, erase your wrong answers completely, and then mark your sheet correctly. In the test on the following pages you will have more questions like the ones you have Just done. Answer each question carefully, but if you are not sure, make the best guess that you can. If you should finish before time is called, go back and check your answers. DO NOT. TURN THIS PAGE UNTIL THE SIGNAL "READY - BEGIN" IS GIVEN. RESTRICTED 253 ANALOGIES Test Questions Mark only on separate answer sheet. 61. AIRPLANE Is to air as BOAT is to 1 fish 2 water 3 sail 4 oar 5 captain 62. AIR is to breathe as WATER is to 1 wash 2 swim 3 drink 4 flow 5 dive 63. TRACK is to train as ROAD is to 1 automobile 2 country 3 tractor 4 concrete 5 travel 64. MAN is to food as MOTOR is to 1 automobile 2 gasoline 3 starter 4 grease 5 driver 65. ERASER is to pencil mark as SOAP is to 1 bath 2 dirt 3 water 4 cleanliness 5 skin 66. CREW 1s to men as FLEET Is to 1 destroyers 2 officers 3 seas 4 ships 5 seapower' 67. TACK is to nail as NAIL is to 1 spike 2 Hammer 3 bolt 4 wedge 5 screw 68. SEEP is to gush as SPRINKLE is to 1 drown 2 p’our 3 dip 4 tumble 5 overflow 69. GIANT is to dwarf as BOULDER is to 1 man 2 rock 3 mountain 4 sand 5 pebble 70. SAPLING is to tree as CHILD is to 1 adult 2 growth 3 maturity 4 youth 5 Infant 71. PEAT Is to self-satisfaction as BLUNDER is to 1 silence 2 excuse 3 grumbling 4 embarrassment 5 ridicule 72. BATTLE is to Navy as GAME is to 1 player 2 college 3 skill 4 competition 5 team 73. BLOCKADE is to commerce as TOURNIQUET is to 1 limb 2 wound 3""bleodlng 4 first aid 5 bandages 74. TRUCK is to utility as LIMOUSINE is to 1 wealth 2 transportation 3 ornamentation 4 expense 5 luxury 75. HAND is to finger as FOOT is to 1 shoe 2 ground 3 ankle 4 leg 5 toe 76. UMBRELLA is to rain as HELMET is to 1 head 2 shrapnel 3 poison gas 4 protection 5 firearms 77. HOUSE is to architect as AUTOMOBILE is to 1 mechanic 2 manufacturer 3 owner 4 designer 5 driver 78. WILDERNESS is to houses as CONTENTMENT is to 1 Joy 2 poverty 3 troubles 4 pleasures 5 duties 79. CUT is to break as EXACT is to 1 distorted 2 mistaken 3 disjoined 4 approximate 5 unskillful DO NOT STOP. GO ON TO THE NEXT PAGE. 254 RESTRICTED 80. WAGES is to work as IMPRISONMENT is to 1 criminal £ Justice 3 crime 4 reform 5 punishment 81. ILLNESS is to symptom as GUILT is to 1 proof 2 evidence 3 misdeed 4 witness 5 conscience 82. EXPERIENCE is to memory as INJURY is to 1 cure 2 salve pain 4 scar 5 revenge 83. POLL TAX is to voting as TARIFF is to 1 government 2 Industry 3 taxation 4 amusements 5 imports 84. WATER is to thirst as SUCCESS is to 1 ambition 2 theory 3 failure 4 ability 5 opportunity 85. PERFORMANCE is to applause as CHARACTER is to 1 respect 2 training 3 actions 4 appearance 5 self-esteem 86. HOE is to weeds as CANNON is to 1 bullets 2 battle 3 enemy 4 battlefield 5 destruction 87. TEACHING is to ignorance as FIRE is to 1 cold 2 heat 3 fuel 4 temperature 5 boiler 88. FLUID is to strainer as NEWS is to 1 public 2 opinion 3 analysis 4 censorship 5 newspaper 89. BUILDING is to plans as BOOK is to 1 outline 2 illustrations 3 preface 4 topic 5 review 90. FAMILY is to home as FLEET is to 1 ocean 2 harbor 3 base 4 flagship 5 native land 91. BREAKDOWN is to repairs as DELINQUENCY is to 1 rehabilitation 2Tpunishment 3 amnesty 4 obedience 5 integrity 92. SOW is to reap as PRACTICE is to- 1 exert 2 compete 3 finish 4 prepare 5 excel 93. MATCH is to flame as TALENT is to 1 effort 2 achievement 3 fame 4 training 5 ability 94. SUCCEED is to opportunity as ENTER is to 1 encouragement 2 room 3 key 4 trickery 5 challenge 95. MONTH is to time as INCH is to 1 ruler 2 measure 3 quantity 4 foot 5 length 96. NEATNESS is to disorder as GOVERNMENT is to 1 freedom 2 citizens 3 opposition 4 anarchy 5 fascism 97. FUEL is to fire as ENERGY is to 1 strength 2 inspiration 3 activity 4 health 5 purpose 98. PRAIRIE is to garden as FOREST is to 1 trees 2 Jungle 3 lumber 4 field 5 orchard 99. FABRIC is to thread as KNOWLEDGE is to 1 fact 2 intelligence 3 Instruction 4 belief 5 Judgment 100. CHICK is to omelette as CALF is to 1 beef 2_hay S cream 4 cheese 5 shoes STOP. WAIT FOR FURTHER INSTRUCTIONS. RESTRICTED 255 UNITED STATES NAVY GENERAL CLASSIFICATION TEST 256 RESTRICTED UNITED STATES NAVY TESTS OP READING AND ARITHMETICAL REASONING FORM I H A V P E R S 16510 1943 NOTICE; This test is not to be shown, or the contents revealed, to unauthorized persons in or out of the Navy, or reproduced in whole or in part without written authorization from the Bureau of Naval Personnel of the United States Navy. ITiis test booklet must be surrendered with the answer sheet at the dose of the test. DO NOT OPEN THIS BOOKLET UNTIL YOU ARE TOLD TO DO SO, RESTRICTED TEST I READING Directions and Samples In this test you will be given several paragraphs to read. Following each para- graph are a number of incomplete sentences, each of which is followed by five phrases lettered from (a) to (e). One of the five phrases added to the in- complete sentence makes a statement which according to the paragraph is true. After reading the paragraph, you are to decide which phrase makes each statement true. Then black in the space on the separate answer sheet which corresponds to that phrase. Read the following sample paragraph; then answer question A. After a can of paint has been opened and the paint partly used, the can should be covered and kept as air-tight as possible to prevent a paint scum from forming on the surface. If scum forms, the paint should be strained through a fine-mesh wire screen or cheesecloth. A. To prevent scum from forming in a partly-used can of paint, one should (a) keep the paint free from dirt. (b) fill the can up with water. (o) stir the paint well before storing, (d) keep the can tightly covered. (e) make sure that the can is more than half full. The paragraph tells you that the can should be covered and kept as air-tight as possible to prevent scum from forming; so (d) is the phrase which completes the statement correctly. Therefore (d) has been blacked in on the separate answer sheet opposite A in the section labeled "Reading — Samples". Now answer question B and black in the space in line B on the answer sheet which corresponds to the pErase which completes the statement correctly, according to the paragraph. B. The method of removing scum from paint is (a) to stir the paint thoroughly. (b) to pour the scum off. (c) to strain the paint through a wire screen. (d) to skim the scum off with a putty knife. (©) not discussed in the paragraph. DO NOT STOP. GO ON TO THE NEXT PAGE. 258 RESTRICTED Answer question C and black in the space on the answer sheet which corresponds to the correct pErase. C. The method of thinning paint is (a) to mix it with turpentine. (b) to mix it with linseed oil. (c) to mix it with whits lead. (d) to mix it with water. (e) not discussed in the paragraph. Now check your answers while I give the correct ones to you. ARE THERE ANY QUESTIONS? When the signal is given, you will have more paragraphs to read and questions to answer. Hark each answer in the proper space on the separate answer sheet. Mark only one answer to each question. Do not make any marks in this booklet. Work as fast and as accurately as you can. Answer every question. If you are not sure of the answer, make the best guess you can. DO NOT TURN THIS PAGE UNTIL THE SIGNAL "READY BEGIN" IS GIVEN RESTRICTED 259 READING Teat Questions Mark only on separate answer sheet. Cruisers have light armor, carry guns of moderate size, and are able to travel at high speed* Cruisers whose largest guns are greater than 6 inches are known as heavy cruisers, while those whose largest guns are 6 Inches or less are known as light cruisers. All cruisers have very large fuel tanks in order to maintain high speed for a long time, A cruiser is divided into numerous water-tight compartments, so that a hole in one part of the ship will flood only a part of the ship. 1, Cruisers whose largest guns are 6-inch guns are (a) light cruisers. (b) medium cruisers, (c) heavy cruisers. (d) auxiliaries, (e) not described in the paragraph. 2, Cruisers have (a) heavy armor, high speed, and large guns, (b) light armor, low speed, and medium guns, (c) light armor, high speed, and small guns, (d) heavy armor, high speed, and medium guns, (e) light armor, high speed, and medium guns. 3, Cruisers are able to maintain high speed for a long time because (a) they have a wide beam, (b) they are divided into many small air-filled compartments. (c) they have only three gun turrets. (d) they carry a large supply of fuel oil. (e) they do not have armor plate. 4. In order to make it more difficult to sink cruisers, they are (a) equipped with thick armor, (b) armed with large guns, (c) fitted with mine-laying equipment. (d) powered by turbines with reduction gear. (e) divided into sections that can be flooded without affecting other parts. DO NOT STOP. GO ON TO THE NEXT PAGE. 260 RESTRICTED Aft-er the setting-up exercises, the men are dismissed and then they must get into uniform for the drill which 1s to follow. Each man should have a uniform handy, either in his hag or locker, so that he can change into the proper uniform for drill in a few minutes. When drill call is sounded, men go quickly to their stations for drill and keep silence, A slow and noisy division is always inefficient. Drills are held to train all hands into an efficient fighting unit. All drills are for this one purpose. If you fail to act in drill as you should in an emergency, you will not know what to do when an emergency comes, 5. When you hear the drill call, you should (a) report to your officer. (b) put on your drill uniform. (c) remain where you are and keep silence. (d) find out where the drill is to be. (e) go at once to your station. 6. Which one of the following is not mentioned in the paragraph? (a) fighting. (b) punishment, (c) exercises, (d) drills. (e) uniforms, 7. If a division is slow and noisy (a) it will be dismissed. (b) it is not neatly dressed. (c) it will behave better in an emergency. (d) it cannot drill well, (e) the men will be punished. 8. The only thing the paragraph tells you about the clothing you wear at setting-up exercises is that (a) it is not drill uniform, (b) it should always be clean and neat. (c) you should be able to find it quickly, (d) it allows you to move freely. (e) it is kept in a locker, 9. The most important thing about your drills is that they (a) make you more alert. (b) make you strong. (c) train you for battle. (d) are required for all enlisted men. (e) teach you everything you should know. DO NOT STOP. GO OH TO THE NEXT PAGE. RESTRICTED 261 If you neglect to obey an order or refuse to obey It, your chances of being sustained are less than one in a thousand. If an officer or petty officer orders you to do a certain thing, and before It Is done, another officer or petty officer orders you to do something else. It Is your duty to inform the officer or petty officer giving you the second order that you have had previous orders, telling him what the orders are and who gave them to you. The officer or petty officer who gave you the second order will then decide whether you are to carry out the first orders or the second orders. 10, The paragraph talks mainly about (a) the importance of good judgment, (b) how officers give orders, (c) respect for officers, (d) obedience to orders, (a) self-reliance. 11, There are specific instructions in the paragraph on (a) courteous behavior to officers, (b) what to do if you receive conflicting orders, (c) how to carry out an order. (d) the decisions you will be expected to make, (o) how to decide whether the first or second officer is right. 12, Should you over tell an officer giving a second order that it conflicts with previous orders? (a) yes, if it does not interfere with the immediate following out of orders which have been given by another officer, (b) only if he has a lower rank than the officer giving the first order. (c) only if the officer asks you, (d) the paragraph does not answer this question, (e) yes, you should always do this when you receive conflicting orders. 13, You will be justified if you (a) occasionally do not obey an order. (b) disobey an order only once in a thousand times, (c) decide not to obey an order because you have received a second order before you carried out the first, (d) always obey the second of two orders, (e) neglect to obey an order because another officer tells you not to obey. 14, Your responsibility to the officer who gives the first of two successive orders (a) is maintained until his order is carried out, (b) includes making him a report of the conflicting order, (c) is not specifically mentioned in this paragraph, (d) is discharged at the moment when a second order is received, (e) cannot be transferred to the officer issuing the second order. DO NOT STOP. GO ON TO THE NEXT PAGE. 262 RESTRICTED The highest deck extending from bow to stern is called the main deck, A partial deck above the main deck at the bow only and not covering amidships Is called the forecastle deck. A partial deck at the stern only and not covering amidships Is called the poop deck, A complete deck below the main deck Is called the second deck. Where there are two or more complete decks below the main deck, they are called second deck, third deck, fourth deck, etc, A partial deck above the lowest complete deck and below the main deck Is called a half deck. Decks which for protective purposes are fitted with plating of extra strength and thickness are defined for technical purposes as protective and splinter In addition to their regular names* Where there is only one such deck. It Is defined as protective, and where there are two, that having the thicker plating Is defined as protective, and that having the thinner plating is defined as splinter. In addition to the regular names* 15* The second deck Is (a) above the main deck. (b) below the main deck. (c) not defined in this paragraph, (d) not a complete deck. (e) so-called because It Is protective. 16, A partial dock above the main deck covering all except the bow Is (a) called the upper deck. (b) not discussed in this paragraph, (c) called the poop deck. (d) called the second deck. (e) always a protective deck. 17. A splinter deck Is (a) not discussed in this paragraph. (b) not for protective purposes* (c) a dock with extra plating but leas heavily armored than the protective (d) the same as the second deck. (e) a half deck which Is usually but not always heavily armored and which deck. Is above the main deck. 18, The main deck Is (a) always armor-plated. (b) always a splinter deck, (c) below the second deck, (d) the largest partial protective (e) always a complete deck. deck. 19* A deck extending from bow to stern above the main dock (a) Is usually but not always for protective purposes. (b) Is the forecastle deck. (c) does not exist. (d) 1s armor-plated. (e) is called the protective second deck. 20* If there were two decks of extra strength and thickness and that with the thinner plating was a complete deck immediately below the main deck. It would be called the (a) protective poop deck. (b) forecastle deck, (c) splinter half deck. (d) splinter second deck, (e) protective lower deck. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 263 When a magnetized needle is mounted on a horizontal axis so that the angle between the needle and a horizontal plana can be measured, it is called a dipping needle. The amount of dip varies according to the place on the earth where the dip is measured. In other words, the needle acts as though the earth itself were a huge magnet with a north and a south pole. At the north magnetic pole the needle will point directly downward. On a line roughly halfway between the north and south magnetic poles, the needle will be horizontal; this line Is called the magnetic equator. At all other points the needle will be neither horizontal nor vertical, and the amount of dip will depend on the distance from the magnetic poles. However, the north magnetic pole of the earth is not identical with the north geographic pole, but is about 20 degrees south of it, a degree being equal to about 70 miles. When traveling on the earth's surface many points are found where the inclination or dip of the needle is the same. A line drawn through these points on a map is called an isoclinlc line. The isoollnic line drawn through points of zero dip, that is, where the needle is horizontal, is called an aclinic line. 21. The statement that a map of the world Indicating Isocllnlc lines would show only one aclinic line (a) Is false. (b) cannot be made on the basis of Information given above. (c) can be Inferred from Information contained in the paragraph. (d) means that there Is only one point on the earth*s surface where a dipping needle would read zero. (e) Implies that the aclinic line and the geographic equator are the same. 22, A dipping needle at the north geographic pole would (a) point in a horizontal direction, (b) point in a nearly vertical direction. (c) point straight downward, (d) show about the same dip as at the equator. (e) show a small angle with the horizontal. 25, Isoclinlc lines on a map of the world would (a) have a general direction which cannot be even approximately determined from the above paragraph. (b) tend to run north and south, (c) tend to run east and west, (d) be very far apart, (e) cross one another in an unpredictable way. 24, The magnetic equator (a) is the same as the geographic equator, (b) passes through the nprth and south magnetic poles, (c) is an aolinio line. (d) is neither an aclinic nor an isoclinlc line. (e) is not defined in the above paragraph. 25, It can be inferred from the above paragraph that the magnetic and geo- graphic equators (a) will intersect at only one point, (b) are identical. (c) are parallel lines, (d) bear no fixed relationship to each other, (e) will intersect at two points. DO NOT STOP. GO ON TO THE NEXT PAGE. 264 RESTRICTED Electric current which is continually changing its magnitude and revers- ing its direction of flow in a conductor is called alternating current; and may be detected either by measuring the amount of current or by measuring the electro-motive-forco (e.m.f.) exerted by it. When either the alternating current or the e.m.f. has passed from zero to its maximum value In one direction, back to zero, then to ita maximum value in the other direction and back to zero. It is said to have completed a cycle. The time required for the current or e.m.f. to pass through one complete cycle is called a period. When the periods are all of the same length, the current produced is called periodic current. If the current goes through the complete cycle 60 times per second, it is said to have a frequency of 60 cycles. Each complete cycle may be thought of as con- taining or measuring 360 degrees. A half-cycle or the change from zero to max- imum In either direction and back to zero would be 180 degrees. If the current reaches a maximum before the e.m.f,, the current is said to be leading the e.m.f. In phase. If the current reaches a maximum after the e.m.f., the current is said to lag In phase. In either case, they are said to be out of phase. 26. A current Is not called alternating unless (a) reversals of direction of flow are Instantaneous. (b) its changes in size or amount are large. (c) its magnitude is always changing. (d) its frequency is 60 cycles per second. (e) it is in phase with its e.m.f. 27. Current which is continually reversing its direction of flow but which requires a different length of time to complete each cycle (a) is called periodic alternating current, (b) is not alternating current, (c) does not exist. (d) is alternating but not periodic current. (e) is called direct current. 28. If a periodic alternating current has a frequency of 100 cycles, the period for this current (a) will be 1/100 of a second* (b) cannot be determined without additional information. (c) cannot be Inferred solely from information given in the paragraph. (d) will be one hundred seconds. (e) is sometimes, but not necessarily, constant. 29. If the current and e.m.f. in a conductor are in such a relationship that one reaches a maximum in one direction when the other is at a maximum in the opposite direction, then (a) the current and e.m.f. are said to be in phase. (b; it can be concluded that no current will flow in the conductor, (c) they can be said to be 90 degrees apart in phase. (d) the current is leading the e.m.f, by three-fourths of a cycle. (e) the current and e.m.f. are 180 degrees apart in phase. 30. The statement that one-half of a cycle is called an alternation (a) is not true. (b) is not to bo derived from the above paragraph. (o) means that the change in one direction is always equal in amount to the change in the opposite direction. (d) could be inferred from statements made In the paragraph, (e) means that the number of cycles per second is equal to the frequency. STOP. WAIT FOR FURTHER INSTRUCTION. RESTRICTED 265 TEST II ARITHMETICAL REASONING Directions and Samples In this teat you will be given some problems in arithmetic* Following each problem are five answers, lettered from (a) to (e); one of these answers is correct* Your task is to solve each problem and black in the space on the separate answer sheet which corresponds to the answer you think is correct* The following problems are samples. D* A boy bought a War Stamp every day for 4 days* How much did he invest in Stamps? (a) #0.50 (b) #1.00 (c) #1.50 (d) $2.00 (e) #2.50 The correct answer to the problem is $1.00, which is answer (b). Notice in row D on the separate answer sheet in the section labeled Arithmetical Reasoning —• Samples" that space (b) has already been blacked in. Now work the next problem, and in row E on the separate answer sheet black in the space which corresponds to the correct answer. E. A ball team played 27 games and won 16 of them* How many did it lose? (a) 1 (b) 7 (c) 10 (d) 11 (e) 43 Solve the next problem, and in row F on the answer sheet black in the space which corresponds to the correct answer. P. If an airplane makes a trip of 500 miles In 2 hours, what Is Its average rate in miles per hour? (a) 250 (b) 251 (c) 498 (d) 502 (e) 1,000 Now check your answers as I give the correct ones to you. ARE THERE ANY QUESTIONS? When the signal is given you will work more problems of the same kind. Mark each answer in the proper space on the separate answer sheet. Mark only one answer to each problem. Do not make any marks in this booklet. Work as fast and as accurately as you can. Work every problem. If you are not sure, mark the answer you think is correct. STOP HERE. DO NOT GO TO THE NEXT PAGE UNTIL THE SIGNAL "READY BEGIN" IS GIVEN. RESTRICTED 267 ARITHMETICAL REASONING Test Problems Solve each problem* Then black in the space corresponding to the correct answer* 31. The U. S, Navy proposes to have 700 combatant ships by a given date. If the Navy has 300 of the ships now, how many more will be needed? (a) 300 (b) 400 (c) 200 (d) 500 (e) 1,000 32. During World War I, keels were laid at Hog Island at the rate of one every 5 days. At that rate, how many keels were laid in 275 days? (a) 51 (b) 55 (c) 75 (d) 270 (e) 1375 33. A speed of 6 miles per minute is equivalent to how many miles per hour? (a) 10 (b) 60 (c) 100 (d) 240 (e) 360 34. A $50 war bond costs $37,50, How many of these bonds may be bought for $150? (a) 1 (b) 2 (c) 3 (d) 4 (e) 5 35, X*s bank account is overdrawn $20.50, How much must he deposit to have a balance of $100 in his favor in the bank? (a) $79.50 (b) $80.50 (c) $119.50 (d) $120.50 (e) $125.00 36, A train leaving New York at 11:15 A.M, is scheduled to make the trip to Philadelphia in 1 hour and 53 minutes. When is it due to arrive? (a) 12:08 P.M. (b) 12:38 P.M. (c) 12:41 P.M. (d) 1:03 P.M. (e) 1:08 P.M. 37, What will be the cost for gasoline on a trip of 180 miles if the car averages 15 miles per gallon and gasoline costs 20 cents a gallon? (a) $1.35 (b) $1.67 (c) $2.40 (d) $2.50 (e) $54.00 38, London and Cologne, 330 miles apart by air, will be how many inches apart on a map whose scale is 50 miles to one inch? (a) *15 (b) 3.1 (c) 5 (d) 6.6 (e) 16.50 39, A "plebe" makes $780 a year at Annapolis. If he receives $300 Immediately, and the rest is paid in 12 equal monthly installments, how much is each monthly installment? (a) $40*00 (h) $45.50 (c) $57.60 (d) $65.00 (e) $90.00 DO NOT STOP. GO ON TO THE NEXT PAGE. 268 RESTRICTED 40. A destroyer travelled for two of three hours at full speed and then reduced its speed by 15 nautical miles during the third hour. If the total distance covered was 81 nautical miles what was its full speed in nautical miles per hour? (a) 22 (b) 27 (o) 30 (d) 32 (e) 35 41. A destroyer division consists of 4 ships. Two divisions plus one leader make a squadron. Pour squadrons make a flotilla. How many ships are there in a flotilla? (a) 20 (b) 32 (c) 33 (d) 36 (e) 40 42, A propeller, with an initial speed of 25 revolutions per second, decreases its speed by 4 revolutions per second each 10 seconds. How many revolutions per second will it be making at the end of 30 seconds? (a) 0 (b) 1 (c) 4 (d) 13 (e) 23| o 43. In an armor-piercing shell Z% of the total weight of the shell is bursting charge. What is the weight, in pounds, of the bursting charge in a 2,100 pound shell? (a) 7 (b) 63 (c) 300 (d) 630 (e) 700 44, A cruiser can fire a total of 165 rounds a minute from its 15 six-inch guns. If six of its guns are "knocked out" how many rounds per minute can be fired? (a) 66 (b) 75 (c) 90 (d) 99 (e) 135 45. If 3 parts sand by weight are mixed with 4 parts cement, what fraction of the total dry mixture is sand? (a) s (b) * (c) 1 (d) » (•) | 46. Carrier A has 80 planes and carrier B has 75 planes. If £ of the of the planes on B are disabled, how many planes are left to protect the carriers? (a) 45 (b) 55 (c) 100 (d) 110 (e) 135 47. If, on the average, 4% of shipped fruit is spoiled, how many boxes .should be shipped if 48 boxes of good fruit are wanted? (a) 4Bj (b) 49 (c) 50 (d) 52 (e) 60 48, How many yards of material 36 Inches wide will it take to black out a window measuring 2 yards wide and 3 yards long? (Assume no overlap of material.) (a) 2 (b) 3 (c) 5 (d) 6 (e) 6| DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 269 49. In 1937, 6.2 cents of the railroad dollar went for fuel. What was the fuel cost per thousand dollars Income? (a) |.62 (b) $6.20 (c) $62.00 (d) $161.00 (e) $620.00 50. The number of sailors in a certain station increased from 5 men to 30,000 men in 5 weeks. What was the average increase per week in the number of sailors? (a) 1,200 (b) 5,991 (c) 5,999 (d) 6,000 (e) 6,001 51. If figs contain 78$ water by weight, how much will 50 pounds of fresh figs weigh when dried? (Assume all water is removed.) (a) 10 lbs. (b) 11 lbs. (c) 13 lbs. (d) 39 lbs. (e) 40 lbs. 52. A gap of 1 5/8 inches is to be filled in with four layers of metal. What will be the thickness in Inches of the fourth layer if the other three are respectively 7/8 inch, l/3 inch, and .25 inch thick? (a) 1/3 (b) 1/4 (c) 1/6 (d) 1/8 (e) .25 53. The capacity of a standard coach is 80. If a troop train of 10 standard coaches is 20$ overloaded, how many more cars would be needed to accommodate all the troops without overloading? (a) 1.6 (b) 2 (c) 2.5 (d) 4 (e) 6 54. The approximate weight in pounds of a shell is equal to one-half the cube of the diameter (measured in inches) of the gun. What must the diameter of a gun be to fire a shell weighing 32 pounds? (a) 2 (b) 4 in. (c) 4 (d) 8 in. (e) 2l| in. 55. Water pressure P (in pounds per square foot) varies directly with depth h (measured in feet), according to the fonnula P • hDg, where g = .003. If the density D of water is 64 pounds per cubic foot, what is the pressure in pounds per square foot on a diver 20 feet below the surface? (a) .01 (b) .38 (c) 3.64 (d) 9.6 (e) 104.2 56. Fifty-four ships of 5,000 tons each are required to carry 2 British divisions. If a 25,000 ton ship holds 5 times as much as a 5,000 ton ship, how many 25,000 ton ships are needed to carry 15 British divisions? (a) 36 (b) 81 (o) 324 (d) 405 (e) 648 57. A plane makes a trip of 1,800 miles in 6 hours. How many hours would it have taken If the speed had been increased by 1 ? 3 (a) 4 (b) 4i (c) 8 (d) 3 (e) 3j DO NOT STOP. GO ON TO THE NEXT PAGE. 270 RESTRICTED 58. A patrol bomber flies due south from its base for 2 hours at 300 miles per hour and then due east for 4 hours at 200 miles per hour. At how many miles per hour must he fly in going directly back to base to reach there in 5 hours? (a) 200 (b) 250 (c) 250Vl0 (d) 266 2/3 (e) 280 59. Countries X, Y, and Z were to be allotted all planes produced in November in the ratio 2:3:5. On that plan the quota for country X was 60 planes. If the allotment ratio for these planes were changed to 1:1:1, how many more planes would X receive? (a) 60 (b) 20 (c) 30 (d) 40 (o) 90 60. Ship A leaves port at 12:00 noon and approaches position X, 240 miles away, at 12 miles per hour. Ship B, 1080 miles from X, leaves 2 hours later. At how many miles per hour must ship B travel in order to reach position X at the same time as ship A? (a) 49.1. (b) 54 (c) 60 (d) 63 (e) 81 11 STOP. WAIT FOR FURTHER INSTRUCTIONS. RESTRICTED 271 UNITED STATES NAVY TESTS OP READING AND ARITHMETICAL REASONING FORM 2 N A V P E R S-16512 1943 NOTICE; This test la not to be shown, or the contents revealed, to unauthorized persons In or out of the Navy, or reproduced In whole or in part without written authorization from the Bureau of Naval Personnel of the United States Navy. This test book- let must be surrendered with the answer sheet at the close of the test. DO NOT OPEN THIS BOOKLET UNTIL YOU ARE TOLD TO DO SO. RESTRICTED 273 TEST I READING Directions and Samples In this test you will be given several paragraphs to read. Following each para- graph are a number of Incomplete sentences, each of which Is followed by five phrases lettered from (a) to (e). One of the five phrases added to the In- complete sentence makes a statement which according to the paragraph la true. After reading the paragraph, you are to decide which phrase makes each statement true. Then black In the space on the separate answer sheet which corresponds to that phrase. Read the following sample paragraph; then answer question A, After a can of paint has been opened and the paint partly used, the can should be covered and kept as alr-tlght as possible to prevent a paint scum from forming on the surface. If scum forms, the paint should be strained through a flne-mesh wire screen or cheesecloth. A. To prevent scum from forming In a partly-used can of paint, one should (a) keep the paint free from dirt. (b) fill the can up with water. (c) stir the paint well before storing. (d) keep the can tightly covered. (e) make sure that the can Is more than half full. The paragraph tolls you that the can should be covered and kept as alr-tlght as possible to prevent scum from forming; so (d) Is the phrase which completes the statement correctly. Therefore (d) has been blacked In on the separate answer sheet opposite A In the section labeled ’’Reading — Samples". Now answer question B and black in the space in line B on the answer sheet which corresponds to the phrase which completes the statement correctly, according to the paragraph. B, The method of removing scum from paint 1s (a) to stir the paint thoroughly, (b) to pour the scum off, (c) to strain the paint through a wire screen, (d) to skim the scum off with a putty knife. (e) not discussed in the paragraph. DO NOT STOP. GO ON TO THE NEXT PAGE. 274 RESTRICTED Answer question C and black in the space on the answer sheet which corresponds to the correct phrase. C. The method of thinning paint is (a) to mix it with turpentine. (b) to mix it with linseed oil, (c) to mix it with white lead. (d) to mix it with water, (e) not discussed in the paragraph Now check your answers while I give the correct ones to you. ARE THERE ANY QUESTIONS? When the signal is given, you will have more paragraphs to read and questions to answer, Mark each answer in the proper space on the separate answer sheet. Mark only one answer to each question. Do not make any marks in this booklet. Work as fast and as accurately as you can. Answer every question. If you are not sure of the answer, make the best guess you can. DO NOT TURN THIS PAGE UNTIL THE SIGNAL "READY - BEGIN" IS GIVEN. RESTRICTED 275 READING Test Questions Mark only on separate answer sheet. Success in battle, the primary aim in every military organization, re- quires implicit obedience to orders. It requires that men be trained to do habitually everything that must be done in battle when under the fire of the enemy. Briefly, discipline is the habit of obedience by which a man obeys an order naturally and without question, without stopping to consider whether he wants to obey it or not; he must learn to obey simply because the order comes from higher authority. 1. Orders must be obeyed whenever (a) it is the sensible thing to do. (b) obedience protects the safety of others. (c) they are reasonable, (d) speed is essential. (e) they come from a person of higher rank. 2, The subject of this paragraph is (a) the speed with which a man must learn to do his Job. (b) the way in which orders are given. (c) tlie necessity of doing Immediately what you are told to do, (d) the consequences of disobedience, (e) the form of a military organization. 3. In obeying an order. It is important to (a) disregard your own immediate wishes, (b) consider carefully whether the order should be obeyed (c) do it deliberately and cautiously. (d) know that the order was proper. (e) consider whether it Involves danger to others. 4. The final object of requiring instant obedience to orders is to (a) avoid argument. (b) maintain authority in the hands of the officers, (c) win battles, (d) enable men to get along with one another. (e) maintain military discipline. DO NOT STOP. GO ON TO THE NEXT PAGE. 276 RESTRICTED An Important task on a naval vessel is to supervise the firing of guns. This is the task of the chief fire-control officer (G.P.C.O.), The C.P.C.O, is in command of all subdivisions of the armament, but he normally performs only supervisory functions in connection with all but the main battery; the latter he directly controls. The fire-control tower is the C.P.C.O.'s battle station. This station is equipped with the necessary fire-control instruments for directing and observing the fire of the main battery. It la also equipped with the necessary communications so that the-C.P.C.O, may at all times com- municate directly with the captain, with the officers in immediate control of the secondary, antiaircraft, and torpedo batteries, and with the officers in charge of subdivisions of the main battery, such as plotting-room officer, turret officers, and spotters. The C.P.C.O. keeps the captain informed as to the range and practicability of opening fire. He gives the commands "commence firing" and "cease firing," when directed by the captain. Prior to opening fire with the main battery he designates the fire distribution to be employed, the targets to be fired at, and the gun directors to be used. He supervises the salvo signals of the main battery and determines the rate of fire. 5, This paragraph concerns mainly (a) the number of batteries that should be fired at a target. (b) the purpose of equipment in the fire-control tower, (c) the knowledge required by the C.P.C.O, (d) the function of the captain -in time of battle. (e) the duties of th© C.P.C.O. 6. The relationship between the C.P.C.O. and the captain is such that (a) it would be necessary for the captain to communicate with the C.P.C.O,, but not the reverse. (b) two-way communication between C.P.C.O, and captain is absolutely necessary, (c) the captain gives the commands to commence and cease firing upon the advice of the C.P.C.O, (d) the captain determines the distribution of fire of the main battery, (e) the captain and C.P.C.O, have different duties so that one does not need to confer with the other. 7, The correct description of what the C.P.C.O, does is that he (a) is in direct control of all the armament. (b) commands the ship during battle, (c) directs the handling of the ship. (d) supervises the use of all of the ship's armament in battle. (e) operates the instruments used to observe the results of fire. 8. One of the main duties of the C.P.C.O, ds to (a) name the targets to be fired at, (b) control the secondary battery fire, (c) operate the gun directors, (d) control the speed of the ship during battle, (e) plot the ship's course. 9* The fire-control tower is evidently a kind of central office for (a) the plotting of the ship’s course. (b) the coordination of information regarding targets and available gun-power. (c) the direction of secondary battery fire only. (d) the spotting of the results of the fire. (e) officers in charge of the various batteries so that they may direct their batteries from it. DO NOT STOP. GO ON TO THE NEXT PAGE. RESTRICTED 277 The one great danger in a small boat when manning before a surf (that Is, traveling In the same direction as the breaking waves! Is that of broachlng-to, which means being thrown broadslde-on In the surf. A boat broachos-to when running before the sea because she opposes little resistance to It and hence Is carried before the wave. Broachlng-to Is most likely to occur when a boat Is headed for shore, because the incoming waves or rollers, on reaching the boat, throw up the stern and as a consequence depress the bow. If she then has suffi- cient Inertia (the greater the weight, the greater the Inertia) to allow the wave to pass her, she will In succession pass through the descending, the horizontal, and the ascending positions as the crest of the wave passes successively her stern, her midships, and her bow. But If a boat on being caught by a heavy wave has not sufficient inertia to allow It to pass her, only the first of the three positions above described occurs. The stern is raised high in the air and the wave carries the boat before It on Its front, sometimes with frightful velocity, and the boat may be thrown end-over-end or be turned broadside to the sea and capsized. 10, In running before a sea, the part of the boat which is lifted first as a wave comes along (a) is the stern. (b) is the bow, (c) is the middle section, (d) can not be determined from the paragraph. (o) would depend on the size of the wave. 11. According to the paragraph, a very light boat would (a) have great inertia because it would ride over the water and not be carried before it, (b) show little danger of capsizing when running through surf, (c) be difficult to handle in a calm sea. (d) have a large inertia because inertia is proportional to the square of the weight, (e) have very little inertia. 12, In a boat running before the sea, there is danger of a boat’s broaching-to when (a) the boat is very heavy, (b) each wave passes the boat quickly by, (c) the waves are high but not breaking. (d) the wave pushes the boat ahead of it, (e) the sea is very deep. 13, One result of an increase in the weight of the boat is to (a) increase the danger of broaching-to, (b) cause the boat to be affected to a greater extent by the wind. (c) make the boat more likely to sink. (d) make the boat more responsive to the force of the waves. (e) make it less likely that the boat will broach-to. 14. The chief topic of the paragraph is (a) the effect of the weight of a boat on its maneuverability. (b) the method of preventing a boat from broaching-to. (c) the way to land a boat through heavy surf. (d) the action of heavy surf on a boat. (e) the power of heavy seas. 15. Since a light boat running before a broken sea opposes little resistance to It, (a) the boat can travel faster than the breaking waves, (b) the waves are likely to break over the bow of the boat# (c) the waves are likely to break over the stern of the boat, (d) the boat may bo pushed along rapidly by the front of the wave. (e) the boat travels so much more slowly than the waves that it is likely to capsize. DO NOT STOP. GO ON TO THE NEXT PAGE. 278 RESTRICTED Articles such as guns, furniture, and special tools which are part of the ship's equipage or which make the ship manageable, habitable, and service- able as a naval vessel are called tltle-B articles. Lost or missing tltle-B articles are surveyed as soon as possible by an officer appointed by the commanding officer If the value of any one article does not exceed #100, or the total of identical articles does not exceed $100, and are surveyed by a board of three officers if values are greater. The survey officer or board makes a searching and exhaustive investigation of the circumstances and In every case determines the cause of the loss. When the loss is due to carelessness on the part of any person, the commanding officer will assign punishment as the case demands. Surveying officers or boards do not hold men responsible for loss due to casualties or stress of weather If witnesses can bo found to prove the cause of loss. 16. One situation mentioned In the paragraph In which men would not be held responsible for loss Is when (a) the loss is due to negligence. (b) the loss Is due to theft. (c) several men are responsible and when witnesses cannot be found to prove responsibility. (d) the loss is due to heavy seas if witnesses testify to this. (e) the loss is due to unavoidable deterioration through use. 17, Lost or missing tltle-B articles not exceeding $100 in value are surveyed by (a) a Chief Petty Officer* (b) some officer with the rank of Lieutenant Commander or higher, who is appointed by the commanding officer. (c) a board of three officers, (d) an officer appointed by the commanding officer. (e) a board consisting of the commanding officer and tho junior officers. 18. The paragraph states that a survey board Investigates the loss of (a) all articles missing from the ship. (b) tltle-B articles which are valued at less than $100. (c) all tltle-B articles. (d) title-B articles valued at more than $100. (e) only those title-B articles lost due to negligence* 19. When the loss of title-B articles has been established as due to careless- ness. (a) punishment is determined by the survey officer or board. (b) punishment usually consists of loss of pay. (c) punishment Is assigned which Is not described in the above paragraph. (d) articles are replaced by the man causing the loss* (e) articles are replaced only If valued at less than $100. 20. The reason for the loss of tltle-B articles Is (a) established only when an individual Is responsible. (b) always determined by a board of officers. (c) always established no matter what the circumstances. (d) rarely successfully established. (e) not established when due to stress of weather* DO NOT STOP, GO ON TO THE NEXT PAGE, RESTRICTED 279 Even during the hottest periods of the year, the temperature rise of the water on the ocean surface is very small compared with the rise in temperature of the land under similar circumstances. There are several reasons why this is true. Except when the sun is nearly overhead, most of its radiation to the surface of the sea is reflected back into space. As the sun approaches the meridian, at which time it is directly overhead, less of its radiation is reflected and more of it is absorbed by the ocean water. However, the sun’s radiation penetrates to a consider- able depth before it is completely absorbed. Yfhereas on land only a few Inches of the surface absorbs all the radiation and has its temperature considerably increased on the sea several fathoms of water absorb the radiation and the temperature in- crease of the water is very small. Another factor which acts to prevent temperature rise of the ocean surface is that some of the heat received is used in evaporating water. The evaporation cools the water and Increases the concentration of salt, both of which increase the density of the water. The denser water then sinks and is replaced by cooler water from below. 21, The reason for the fact that the sea water washing the coast is warmer in summer than in winter is (a) that the temperature change of the surface of the ocean is very small, (b) not discussed in the paragraph, (c) that the land warms the water. (d) that the ocean is very salty, (e) not known to those who have studied the way in which the temperature of the sea changes. 22. Evaporation of sea water (a) acts to raise the temperature of the water. (b) does not affect the rate of heating the water, (c) is so small as to be negligible. (d) occurs more rapidly at night than during the day, (e) tends to counteract the heating effect of the sun’s rays. 23. It can be concluded from statements made in the paragraph that (a) ocean temperatures are in general more stable than land temperatures, (b) land, being solid, retains heat longer and therefore its temperature is more stable than that of water. (cj because of the existence of currents in the ocean its temperature is less stable than land temperature. (d) the surface of the ocean heats more rapidly during the day than does the surface of the land. (e) the prediction of dally variations in ocean temperature is very difficult. 24, The formation of a layer of warm water tends to be prevented by (a) an increase in the density of surface water, (b) absorption of radiant energy which strikes the water at right angles, (c) absorption of heat from the air above the water, (d) the fact that water does not conduct heat well, (e) the fact that water is relatively incompressible. 25. One factor affecting the stability of land temperature as compared to that of sea water temperature is (a) the relative density of land and water. (b) seasonal variation in the concentration of salt in sea water. (c) the fact that the angle at which the rays strike the surface does not change on land as it does on water. (d) the fact that more solar radiation strikes the water than the land. (e) the fact that solar radiation penetrates water more deeply than it does land. DO NOT STOP. 00 ON TO THE NEXT PAGE, 280 RESTRICTED Viewed from a position behind the ship, a right-handed propeller turns clockwise when driving the ship ahead and a left-handed propeller turns counter- clockwise. Reversing the direction of rotation of the propeller drives the ship backward. When turning, the propeller exerts a sidewise force on the propeller shaft because the propeller blades pass through water under a greater pressure at the lower part of their arc than at the upper part of the arc. Thus there is a resultant force against the propeller shaft (and there- fore the stern of the ship) opposite to the direction in which the lower blades travel. Right-handed propeller viewed from rear. Propeller turning clockwise. Ship moving ahead. Direction of rotation. 26, In a ship with a single propeller, the relationship between the sidewise force exerted by a right-handed propeller and the force exerted by a rudder turned to the right (a) is such that the two forces would exactly counteract one another, (b) is such that the force exerted by the rudder would be greater than that exerted by the propeller, (c) cannot be determined from information given in the paragraph. (d) is such that the ship would move to the right, (e) would depend upon whether the ship is turning to the right or left. 27. Suppose a ship has two propellers, one on each side of the ship. If the sidewise force of one propeller is to counteract that of the other, then (a) the two propellers should always revolve in opposite directions, (b) both propellers should turn in the same direction. (c) it does not matter in what direction the propellers turn, pro- vided they are the same size and turn at the same speed, (d) it does not matter in what direction the propellers turn, pro- vided they turn at the same speed. (e) the direction in which the propellers should turn cannot be determined from the information given in the paragraph. 28. If a right-handed propeller is turning in such a direction as to cause the ship to move backward and no other force except that caused by the propeller is operating, the stern of the ship will (a) tend to be pushed to the right. (b) move directly backward. (c) move first to the right and then to the left. (d) move first to the left and then to the right, (e) tend to go to the left. 29. Assuming the same number of revolutions per minute, an Increase In the diameter of the ship's propeller would (a) Increase the sidewise force it would exert. (b) decrease the sidewise force it would exert. (c) have no effect on the tendency of the propeller to move the stern of the ship sidewise. (d) have an effect which cannot be determined from information given in the paragraph. (e) change the direction of the force exerted by the propeller in a way which would depend on the speed of the ship. 30, A right-handed propeller would, when turning clockwise, tend to (a) cause the ship to turn in a circle to the left. (b) cause the ship to move directly ahead. (c) cause the ship to turn in a circle to the rlgnt. (d) turn the ship in an unpredictable direction. (e) exert a directional force which cannot be determined from information given in the above paragraph. STOP. WAIT FOR FURTHER INSTRUCTIONS, RESTRICTED 281 TEST II ARITHMETICAL REASONING Directions and Samples In this test you will be given some problems in arithmetic. Following each problem are five answers, lettered from (a) to (e); one of these answers is correct. Your task is to solve each problem and black in the space on the separate answer sheet which corresponds to the answer you think is correct. The following problems are samples. D, A boy bought a War Stamp every day for 4 days. How much did he Invest in Stamps, (a) $0.50 (b) $1.00 (c) $1.50 (5) $2.00 (e) $2.50 The correct answer to the problem is $1,00, which is answer (b). Notice in row D on the separate answer sheet in the section labeled "Arithmetical Reasoning -- Samples" that space (b) has already been blacked in. Now work the next problem, and in row E on the separate answer sheet black in the space which corresponds to the correct' answer. E. A ball team played 27 games and won 16 of them. How many did it lose? (a) 1 (b) 7 (c) 10 (d) 11 (e) 43 Solve the next problem, and in row P on the answer sheet black in the space which corresponds to the correct answer. F, If an airplane makes a trip of 600 miles in 2 hours, what is its average rate in miles per hour? (a) 250 (b) 251 (c) 498 (d) 502 (e) 1,000 Now check your answers as I give the correct ones to you. ARE THERE ANY QUESTIONS? When the signal is given you will work more problems of the same kind, Mark each answer in the proper apace on the separate answer sheet. Mark only one answer to each problem. Do not make any marks in this booklet. Work as fast and as accurately as you can. Work every problem. If you are not sure, mark the answer you think is correct. DO NOT TURN THIS PAGE UNTIL THE SIGNAL "READY - BEGIN" IS GIVEN, RESTRICTED 283 ARITHMETICAL REASONING Test Problems Solve each problem. Then blacken the space corresponding to the correct answer. 31. One man out of every 6 from a company of 120 volunteered for scout duty. How many men volunteered? (a) 20 (b) 24 (c) 96 (d) 100 (e) 720 32, A train of 5 standard cars accommodates 400 passengers. If another standard car is added, how many passengers can be accommodated altogether? (a) 480 (b) 500 (c) 515 (d) 525 (e) 2,000 33. If a bomb used for practice purposes contains 5 pounds of black powder, 505 pounds of black powder will be sufficient for how many of these bombs? (a) 101 (b) 250 (c) 255 (d) 500 (e) 2,525 34, To strengthen each long ton of steel, 14 pounds of manganese are needed. How many pounds of manganese are needed to strengthen 26 long tons of steel? (a) (b) (c) 12 (d) 40 (e) 364 35. There are 25 to 60 men in a platoon. A company consists of 3 to 4 platoons. What is the greatest number of men which can make up a company? (a) 75 (b) 149 (c) 170 (d) 240 (e) 595 36, If a food ration allows a person 7 ounces of bread a day, how many pounds is person allowed per week? (1 pound = 16 ounces) (a) 1 (b) (c) 3| (d) 16 (e) 784 37. A pace is feet. How many paces are necessary to measure 150 feet? (a) 60 (b) 100 (c) 167 (d) 225 (e) 375 38. A canal, known to be 20 feet wide at point X, measures 1 inch wide in an aerial photograph. A military target lying near X measures 2 inches in diameter In the photograph. What is the actual diameter, in feet, of the target? (a) 10 (b) 30 (c) 40 (d) 80 (e) 40TT 39. If 7 of a ton of coal per day is required to heat a building, how many days 5 will 20 tons last? (a) 4 (b) 10 (c) 25 (d) 75 (e) 100 DO NOT STOP, GO ON TO THE NEXT PAGE. 284 RESTRICTED 40. In a certain electric company a worker's yearly pay has risen from $2,000 to $4,872 in the past 10 years. What 1s the average yearly increase? (a) $14.36 (b) $28.72 (c) $287.20 (d) $687.20 (e) $1,436.00 41. If a mechanized division, averaging 30 miles per hour, leaves its base at 2:10 P.M., how many miles will it have gone by 5:30 P.M.? (a) 9 (b) 39 (c) 96 (d) 100 (e) 200 42, Two towns, 15 miles apart, are inches apart on a map. On that map, one mile is represented by how many inches? (a) (b) (c) - (d) 11- (e) 20 43. The weight of the container is 6% of the shipping weight of a certain food- stuff. How many tons of the foodstuff can be shipped if the total shipping weight is to be 500 tons? (a) 200 (b) 470 (c) 530 (d) 531.9 (e) 8,333 44. The rate of absenteeism has doubled since the start of the war. If 6 out of every 200 workers were absent before the war, what percentage of the workers are absent now? (a) l|£ (b) 6% (c) 16|# (d) % (e) 94% 45. The gasoline consumption .of a car at 60 miles per hour is double the con- sumption at 30 miles per hour. A motorist, traveling at 30 miles per hour, gets 20 miles on a gallon of gasoline. If gasoline costs 20 cents a gallon, how much does he save on gasoline for a 100 mile trip by traveling at the slower speed? (a) $.50 (b) $1.00 (c) $2.00 (d) $3.00 (e) $4.00 46. The United States force at the Solomons consisted of 2 heavy cruisers, 3 light cruisers, and 8 destroyers. What fraction of the force was destroyers? (a) (b)f (c)l (4) | (•)£ 47. Bomber X leaves London for Berlin one-half hour later than bomber Y, How many miles per horn* must X travel to reach Berlin at the same time as Y if Y travels at 200 miles per hour and makes the trip in 3 hours? (a) 66| (b) 166| (c) 17l| (d) 240 (e) 267 48. Pre-enrollment examinations at a certain school eliminate 40% of the candidates for this school. Out of 150 candidates in 1941, how many were accepted into the school? (a) 60 (h) 90 (c) 110 (d) 144 £ (d) | <•>? 57. The time in seconds, t, required by an object to fall d feet is represented 2 by the formula: d * 16t . What is the altitude, in feet, of an airplane if a bomb dropped from it takes 20 seconds to land? (a) (b) 32V§" (c) 320 (d) 5,120 (e) 6,400 DO NOT STOP. 00 ON TO THE NEXT PAGE. 286 RESTRICTED 58. Prom each bonus payment must be deducted 1% for Social Security, A% for state tax, and 5% for Victory tax. How much bonus must an employee be paid if he is to receive $10.00 after deductions have been made? (a) $9,00 (b) $11.00 (c) $11.11 (d) $19.00 (e) $100.00 59. The sides, top, and bottom of a shipping crate are each j- inch thick. If each outer edge of the crate is 1 foot long, what is the inner capacity of the crate in cubic inches? (a) 121 (b) 1,296 (c) 1,331 (d) l,52o| (e) 1,728 60. A patrol plane sights a motionless target 6 miles due south and radios a battleship which Is 5 miles due east of the plane. For what range, in miles, should the guns of the battleship be set to fire on the target? (a) 5| (b) VII (c) V30 (d) V51 (e) 61 STOP. WAIT FOR FURTHER INSTRUCTIONS. RESTRICTED 287 INDEX The subject indexes of all STR volumes are combined in a master index printed in a separate volume. For access to the index volume consult the Army or Navy Agency listed on the reverse of the half-title page. AAF Tactical Center, Orlando, Flor- ida, 65 Ability measurements; see Aptitude tests, types AC (arithmetical computation) test, 21,151 Activity-preference test, 53-57 factor analysis, 55 Kuder preference record, 53 like most-like least form, 54 reliability of scores, 56 reliability of test, 55 scope, 53-55 scoring method, 55 Strong vocational interest blank, 53 Adaptometers, 93-99 clockface adaptometer, 96 Hecht-Schlaer, RCN model, 96 NDRC, Model III, 93-96 Purkinje test, 96 radium plaque adaptometer, 4, 94, 96, 98 Rostenberg, 93, 96 Tufts-SDS test, 96 visual motor coordination test, 96 AGCT (Army general classification) test, 33, 61, 64, 70 AI radar operator tests (airborne interception), 66 Airborne oscilloscope operator tests, 65 Amphibious training; personal in- ventory test, 45, 48 selection of crew, 104-105 Aniseikonia, 71 Antiaircraft personnel; see Height finder operators Aptitude testing methods; construc- tion of tests, 134-145 evaluation of tests, 24 need for new tests, 7 periodic analysis of tests, 8 recommendations for future im- provement, 7-9, 24, 31, 106 Aptitude tests, types; activity-pref- erence test, 53-57 airborne oscilloscope reading test, 65 arithmetical computation test, 21, 151 arithmetical reasoning test, 15, 267-271, 283-287 officers, 184-186, 200-202, 216-218 code, 58-62 combat information center test, 63, 66 Eddy test for radio men, 123 electrical knowledge test, 89 general classification test, 10-25, 45, 150 GIC radar operator test, 65 gunnery proficiency tests, 128 mechanical aptitude test, 11, 72, 89, 150 mechanical comprehension test, 70, 175-183, 191-199, 207-215 mechanical knowledge test, 11, 19, 89, 150 Navy basic classification test bat- tery, 1, 10-26 Navy selective examination, 27 officer qualification test, 1, 27-32, 153, 173-218 oral trade questions, 126 oscilloscope operator tests, 64-67 Purkinje test for night vision, 96 radar, 63-68 radio technician selection test, 126 range and height finder operators, 69-92 Signal Corps code aptitude test, 58 signalman test, 128 sonar pitch-memory test, 126 summary of, 2 tactical radar aptitude test, 65, 66 telephone talkers, 100-102, 128 Tufts-SDS test for night vision, 96 Winchman and Hatchman test, 126 work readiness tests, 107, 131 AR (arithmetical reasoning) test, 11, 15, 28, 150 Navy test, 267-271, 283-287 officer qualification test, 184-186, 200-202, 216-218 Arithmetical computation test, 21, 151 analysis of scores, 21 content, 21 use, 22 validity, 22, 151 Arithmetical reasoning test, 11, 15, 28, 150 Navy test, 267-271, 283-287 officer qualification test, 184-186, 200-202, 216-218 Army general classification test, 33, 61, 64, 70 Bainbridge Naval Training Center; mechanical knowledge test, 18 radio code tests, 61 Battle noise recordings for tests, 49 Bausch and Bomb ortho-rater, 67, 74 Biegel, speed of response test, 58-62 Boca Raton Field, Florida, 66 Botts stereoscopic test, 90 Brown University, 36 California University, 66 Camp Blanding, Fla., 94 Camp Davis, North Carolina, 71 CHL (chain home link) radar operator tests, 65 CIC (Combat Information Center) aptitude test, 63, 66, 67 Classification, mechanical aids, 5 electric accounting card, 119 Findex filing system, 108 interviewing guides, 107 McBee personal record file, 109 Q-card, 108-110 selectometer, 105, 110-117 Classification procedures, 103-133 classification specialists, 105-107, 124, 130-133 coding of occupational skills, 133 evaluation of past experience, 104, 109, 119, 125, 127 for special groups aboard ship, 104, 127 advanced classification, 6, 125- 133 Amphibious Training Com- mand, 104 battleship crews, 104 code operators, 4, 58-62 destroyer crews, 105 fire controlman, 72 landing craft schools, 129 night lookout personnel, 4, 93-99 radar operators, 4, 27, 63-68 range and height finder opera- tors, 4, 69-92 recruits, 5, 118-124 telephone talkers, 4, 100-102 Classification procedures, improve- ment, 118, 121-124, 127-133 interview, 8, 120, 122 job assignment, 120-124, 125-127 RESTRICTED 289 290 INDEX objectives of classification, 5, 103, 118, 125 performance tests, 128-130 permanent classification records, 106 principles of good classification, 105-107 secondary test batteries, 122 specialized ratings, 133 standardized procedure, 106, 130 typical problems, 126 use of service school grades, 130 weighting of personnel informa- tion, 107, 115 Classification test, Army, 33, 61, 64, 70 Classification test battery for Navy, 119, 150 see also Navy classification test Classification tests; see Aptitude tests, types Clockface adaptometer, 96 Coast Artillery Corps, 71 Code aptitude tests, 58-62 code receiving test, 60, 128 low correlation with background, 62 ROA-2; 58 Signal Corps code aptitude test, 58 speed of response test, 58-62 types, 58 validity, 59-61 visual code reception, 62 College Entrance Examination Board; gunnery proficiency tests, 128 Navy officer qualification test, 27 Combat Information Center aptitude test, 63, 66, 67 polar-grid coordinate subtest, 66 predictive ability, 66 relative movement subtest, 66 reliability and validity, 67 scale reading subtest, 66 Coordinate reading test, 65, 66 Cornell selectee index, 49 CRT (code receiving test), 60 Cutoff score, 40, 48 Dartmouth Eye Institute, 71, 77 Dearborn-Johnston test, 69, 81 Destroyer crews, classification, 105 Distractors, 137, 142 Drew Field, Florida, 64 Early Warning radar operator tests, 65 Eddy test for radio men, 123 Eighth Air Force, radar operator selection, 66 Eikonometer, 69-72, 77-79, 87 Electric accounting punched card, 119 Electrical knowledge test, 89 Emotional stability tests, 36-52 Cornell selectee index, 49 New London-NDRC questionnaire, 49 personal inventory, 36-49 use of battle noise, 49 Error, unit of (UOE), 84 Escape tank training, 47 ESF scores (equivalent target area), 98 Esophoria, 70, 72 EW (Early Warning) radar opera- tor tests, 65 Exophoria, 70, 72 Factor analysis, 19-21, 55, 142 Navy classification test, 19-21 principal component method, 55 Findex filing equipment, 108 Fire controlman, selection tests, 72 Fort Eustis, Virginia, 71 Fort Lauderdale, Florida, 72, 88-89 GCI (ground control interception) radar test, 64, 65 General classification test, Army, 33, 61, 64, 70 General classification test, Navy, 10- 26, 45, 150, 233-256 analogies subtest, 11, 15, 240, 252 fleet edition (self-administering), 18 opposites subtest, 11, 15, 237, 249 sentence completion, 10, 15, 234, 246 speed element, 150 Ground control interception radar test, 64, 65 Gunner proficiency tests, 128 Hecht-Schlaer adaptometer, 96 Height finder operators, 4, 69-92, 105 see also Visual tests and instru- ments fire controlman, 72 improved selection criteria, 82-83 performance criterion score, 71, 83 prediction of classroom grades, 89 qualification tests, 70, 90 recommended selection program, 87 validity of performance criteria, 84-87 visual test scores, 88 Horst, item analysis, 152 Hyperphoria, 70, 72 Interocular distance, 73 Interpupillary distance, 70-74, 83, 105 definition, 73 for white and negro populations, 74 Interpupillometer, NDRC, 69, 73 Kuder preference record, 53 LCVP crew, tests and training, 129 Leadership prediction, 1, 32-35, 48 accuracy, 34 Army general classification test, 3 artillery officers, 33 correlation with civilian positions, 3, 33 infantry OCS, 32 Marine Corps OCS, 48 Literacy test, 126 Lookout personnel selection, 4 see also Night vision test devices “Lung,” submarine escape appli- ance, 47 M 2 stereoscopic trainer, 69, 71, 79 MA 2 mechanical comprehension test, 70 MA 3 mechanical comprehension test, 70 Maddox rod spectacles, 74 Massachusetts vision test kit, 69, 71, 74, 76 description, 74 relation to ortho-rater test, 76 reliability, 74 MAT (mechanical aptitude test), 11, 72, 89, 150 McBee Keysort system, 133 McBee personal record file, 109 Mechanical aptitude test (MAT), 11, 72, 89, 150 block counting, 11 mechanical comprehension, 11 speed element, 150 surface development, 11 Mechanical comprehension test, 28, 70 officer qualification test, 175-183, 191-199, 207-215 Mechanical knowledge test, 11, 19, 89, 150 electrical vs. mechanical scores, 19 reliability, 19 RESTRICTED INDEX 291 speed element, 150 statistical analysis, 18 Medical Field Research Laboratory, North Carolina, 48 Medical Research Department, New London, 37 Mirrophones, 101 MKT (mechanical knowledge test), 11, 19, 89, 150 Morse Code, learning process, 58 Multiple-choice tests, advantages, 136 National Research Council Commit- tee on Service Personnel, 37 National Reserve Midshipmen’s School (W.R.), Smith Col- lege, 30 Naval Training School, Dartmouth College, 30 Naval Training School, Fort Lauder- dale, Florida, 66 Naval Training School, Virginia, 65 Naval training stations; Great Lakes, 12 Newport, 12, 39 Norfolk, 12 San Diego, 12 Navy classification tests, new, 15-42 ability measurements, 19, 20, 21, 23 analysis of scores, 17, 19-22 arithmetical reasoning test (AR), 11 characteristics of test, 16 correlation of tests, 17, 23 correlation with service school grades, 24 factor analysis results, 19-21 Form 2; 18 general classification test, 10 mechanical aptitude test (MAT), 11 mechanical knowledge test (MKT), 11 range of difficulty, 15 reading test, 11 recommendations, 24 reliability of tests, 17 validity coefficient, 23 Navy classification tests, old, 11-14 deviation of scores, 14 intercorrelation of tests, 12 item analysis, 12 Navy radio code aptitude test, 11 Navy standard recruit test in arithmetic, 11 Navy standard recruit test in English, 11 Navy standard recruit test in spelling, 11 O’Rourke general classification test, 11 O’Rourke mechanical aptitude test, 11 statistical analysis, 11-14 validity, 13 Navy officer qualification test; see Officer qualification test Na