W^ Volume V: Homicide, Suicide, -„ and Unintentional fid J^Injuries WA : ^00 .. S2r * 5 Report of the Secretary's Task Force on Black & Minority Health U.S. Department of Health and Human Services PROPERTY OF THF NATIONAL yBRARYOF MEDICJNE ILL Volume V: Homicide, Suicide, and Unintentional Injuries Report of the Secretary's Task Force on Black & Minority Health U.S. Department of Health and Human Services January 1986 ..rt v. 5 SECRETARY'S TASK FORCE ON BLACK AND MINORITY HEALTH MEMBERS Thomas E. Malone, Ph.D., Chairperson Katrina W. Johnson, Ph.D., Study Director Wendy Baldwin, Ph.D Betty Lou Dotson, J.D. Manning Feinleib, M.D., Dr.P.H. William T. Friedewald, M.D. Robert Graham, M.D. M. Gene Handelsman Jane E. Henney, M.D. Donald R. Hopkins, M.D. Stephanie Lee-Miller Jaime Manzano J. Michael McGinnis, M.D. Mark Novitch, M.D. Clarice D. Reid, M.D. Everett R. Rhoades, M.D. William A. Robinson, M.D., M.P.H, James L. Scott Robert L. Trachtenberg T. Franklin Williams, M.D. ALTERNATES Shirley P. Bagley, M.S. Claudia Baquet, M.D., M.P.H, Howard M. Bennett Cheryl Damberg, M.P.H. Mary Ann Danello, Ph.D. Jacob Feldman, Ph.D. Marilyn Gaston, M.D. George Hardy, M.D. John H. Kelso James A. Kissko Robert C. Kreuzburg, M.D Barbara J. Lake Patricia L. Mackey, J.D. Delores Parron, Ph.D. Gerald H. Payne, M.D. Caroline I. Reuter Clay Simpson, Jr., Ph.D. Ronald J. Wylie TABLE OF CONTENTS Introduction to the Task Force Report ................ v Members of the Subcommittee.....................ix Report of the Subcommittee on Homicide, Suicide, and Unintentional Injuries ................. 1 Supporting Papers Commissioned by the Subcommittee 1. Robert L. Hampton: Family Violence and Homicide in the Black Community: Are They Linked?................69 2. Darnell F. Hawkins: Longitudinal-Situational Approaches to Understanding Black-on-Black Homicide ............ 97 3. Fred Loya, Philip Garcia, John D. Sullivan, Luis A. Vargas, James Mercy, Nancy Allen: Conditional Risks of Homicide among Anglo, Hispanic, Black, and Asian Victims in Los Angeles, 1970-1979 ...................... 117 4. Fred Loya, Philip Garcia, John D. Sullivan, Luis A. Vargas, Nancy Allen: Changes in the Criminal Homicide Rate of American Indians for the City of Los Angeles: 1970-1979. A Research Note........................137 5. Coramae R. Mann: The Black Female Criminal Homicide Offender in the United States.................145 6, Harold M. Rose: Can We Substantially Lower Homicide Risk in the Nation's Larger Black Communities? ........... 185 7. Deborah Prothrow-Stith: Interdisciplinary Interventions Applicable to Prevention of Interpersonal Violence and Homicide in Black Youth .................... 227 8. Karil S. Klingbeil: Interpersonal Violence: A Comprehensive Model in a Hospital Setting from Policy to Program.......245 9. Burnet B. Sumner, Elizabeth R. Mintz, Patricia L. Brown: Interviewing Persons Hospitalized with Interpersonal Violence-Related Injuries: A Pilot Study ........... 267 ill iv INTRODUCTION TO THE TASK FORCE REPORT Background The Task Force on Black and Minority Health was established by Secretary of Health and Human Services Margaret M. Heckler in response to the striking differences in health status between many minority populations in the United States and the nonminority population. In January 1984, when Secretary Heckler released the annual report of the Nation's health, Health, United States, 1983, she noted that the health and longevity of all Americans have continued to improve, but the prospects for living full and healthy lives were not shared equally by many minority Americans. Mrs. Heckler called attention to the longstanding and persistent burden of death, disease, and disability experienced by those of Black, Hispanic, Native American, and Asian/Pacific Islander heritage in the United States. Among the most striking differentials are the gap of more than 5 years in life expectancy between Blacks and Whites and the infant mortality rate, which for Blacks has continued to be twice that of Whites. While the differences are particularly evident for Blacks, a group for whom information is most accurate, they are clear for Hispanics, Native Americans, and some groups of Asian/Pacific Islanders as well. By creating a special Secretarial Task Force to investigate this grave health discrepancy and by establishing an Office of Minority Health to implement the recommendations of the Task Force, Secretary Heckler has taken significant measures toward developing a coordinated strategy to improve the health status of all minority groups. Dr. Thomas E. Malone, Deputy Director of the National Institutes of Health, was appointed to head the Task Force and 18 senior DHHS executives whose programs affect minority health were selected to serve as primary members of the Task Force. While many DHHS programs significantly benefit minority groups, the formation of this Task Force was unique in that it was the first time that attention was given to an integrated, comprehensive study of minority health concerns. Charge Secretary Heckler charged the Task Force with the following duties: • Study the current health status of Blacks, Hispanics, Native Americans, and Asian/Pacific Islanders. • Review their ability to gain access to and utilize the health care system. • Assess factors contributing to the long-term disparities in health status between the minority and nonminority populations. v m • Review existing DHHS research and service programs relative to minority health. • Recommend strategies to redirect Federal resources and programs to narrow the health differences between minorities and nonminorities. • Suggest strategies by which the public and private sectors can cooperate to bring about improvements in minority health. Approach After initial review of national data, the Task Force adopted a study approach based on the statistical technique of "excess deaths" to define the differences in minority health in relation to nonminority health. This method dramatically demonstrated the number of deaths among minorities that would not have occurred had mortality rates for minorities equalled those of nonminorities. The analysis of excess deaths revealed that six specific health areas accounted for more than 80 percent of the higher annual proportion of minority deaths. These areas are: • Cardiovascular and cerebrovascular diseases • Cancer • Chemical dependency • Diabetes • Homicide, suicide, and unintentional injuries • Infant mortality and low birthweight. Subcommittees were formed to explore why and to what extent these health differences occur and what DHHS can do to reduce the disparity. The subcommittees examined the most recent scientific data available in their specific areas and the physiological, cultural, and societal factors that might contribute to health problems in minority populations. The Task Force also investigated a number of issues that cut across specific health problem areas yet influence the overall health status of minority groups. Among those reviewed were demographic and social characteristics of Blacks, Hispanics, Native Americans, and Asian/Pacific Islanders; minority needs in health information and education; access to health care services by minorities; and an assessment of health professionals available to minority populations. Special analyses of mortality and morbidity data relevant to minority health also were developed for the use of Task Force. Reports on these issues appear in Volume II. Resources More than 40 scientific papers were commissioned to provide recent data and supplementary information to the Task Force and its subcommittees. Much material from the commissioned papers was incorporated into the subcommittee reports; others accompany the full text of the subcommittee reports. VI An inventory of DHHS program efforts in minority health was compiled by the Task Force. It includes descriptions of health care, prevention, and research programs sponsored by DHHS that affect minority populations. This is the first such compilation demonstrating the extensive efforts oriented toward minority health within DHHS. An index listing agencies and program titles appears in Volume I. Volume VIII contains more detailed program descriptions as well as telephone numbers of the offices responsible for the administration of these programs. To supplement its knowledge of minority health issues, the Task Force communicated with individuals and organizations outside the Federal system. Experts in special problem areas such as data analysis, nutrition, or intervention activities presented up-to-date information to the Task Force or the subcommittees. An Hispanic consultant group provided inform- ation on health issues affecting Hispanics. A summary of Hispanic health concerns appears in Volume VIII along with an annotated bibliography of selected Hispanic health issues. Papers developed by an Asian/Pacific Islander consultant group accompany the data development report appearing in Volume II. A nationwide survey of organizations and individuals concerned with minority health issues was conducted. The survey requested opinions about factors influencing health status of minorities, examples of success- ful programs and suggestions for ways DHHS might better address minority health needs. A summary of responses and a complete listing of the organizations participating in the survey is included in Volume VIII. Task Force Report Volume I, the Executive Summary, includes recommendations for department-wide activities to improve minority health status. The recommendations emphasize activities through which DHHS might redirect its resources toward narrowing the disparity between minorities and nonminorities and suggest opportunities for cooperation with nonfederal structures to bring about improvements in minority health. Volume I also contains summaries of the information and data compiled by the Task Force to account for the health status disparity. Volumes II through VIII contain the complete text of the reports prepared by subcommittees and working groups. They provide extensive background information and data analyses that support the findings and intervention strategies proposed by the subcommittees. The reports are excellent reviews of research and should be regarded as state-of-the-art knowledge on problem areas in minority health. Many of the papers commissioned by the Task Force subcommittees accompany the subcommittee report. They should be extremely useful to those who wish to become familiar in greater depth with selected aspects of the issues that the Task Force analyzed. Vll The full Task Force report consists of the following volumes: Volume I: Volume II Volume III: Volume IV: Volume V: Volume VI: Volume VII: Volume VIII: Executive Summary Crosscutting Issues in Minority Health: Perspectives on National Health Data for Minorities Minority and other Health Professionals Serving Minority Communities Minority Access to Health Care Health Education and Information Cancer Cardiovascular and Cerebrovascular Diseases Homicide, Suicide, and Unintentional Injuries Infant Mortality and Low Birthweight Chemical Dependency Diabetes Hispanic Health Issues Survey of Non-Federal Community Inventory of DHHS Program Efforts in Minority Health vm SUBCOMMITTEE ON HOMICIDE, SUICIDE, AND UNINTENTIONAL INJURIES Robert L. Trachtenberg, Chairperson Deputy Administrator Alcohol, Drug Abuse, and Mental Health Administration Thomas L. Lailey Deputy Chief Center for Studies of Antisocial and Violent Behavior National Institute of Mental Health Alcohol, Drug Abuse, and Mental Health Administration Stephanie Lee-Miller Assistant Secretary for Public Affairs Office of the Secretary Department of Health and Human Services Delores L. Parron, Ph.D. Associate Director for Special Populations National Institute of Mental Health Alcohol, Drug Abuse, and Mental Health Administration Clarice D. Reid, M.D. National Coordinator, Sickle Cell Disease Program Chief, Sickle Cell Disease Branch National Heart, Lung, and Blood Institute National Institutes of Health Mark L. Rosenberg, M.D., M.P.P. Chief, Violence Epidemiology Branch Center for Health Promotion and Education Centers for Disease Control IX X REPORT OF THE SUBCOMMITTEE ON HOMICIDE, SUICIDE, AND UNINTENTIONAL INJURIES XI The Subcommittee gratefully acknowledges Marcia R. Feinleib for her valuable aid in preparing this report. CONTENTS Tables and Figures ..... .................... xiv Chapter I. Introduction......................1 Chapter II. Magnitude of the Problem ........ ........ 5 Chapter III. Factors Associated with Excess Deaths from Homicide, Suicide, and Unintentional Injuries .......... 29 Chapter IV. Strategies for Preventive Intervention ......... 41 Chapter V. Subcommittee Recommendations..............61 xiii xiv TABLES AND FIGURES Tables Table 1. Homicide Rates for Selected Countries: 1970-1978 .... 6 Table 2. Death Rates for Homicide and Legal Intervention: 1970-1983 ........................ 8 Table 3. Black Males and Females: Death Rates for 5 Leading Causes of Death, 1981, Selected age groups ....... 9 Table 4. Race of Homicide Victims in the Nation, 1968-1978 .... 12 Table 5. Homicide Rates in 14 Major Cities of the United States.........................15 Table 6. Whites, Chinese, and Japanese: Average Annual Death Rates for Suicide, 1980.................26 Figures Figure 1. Death Rates from Unintentional Injury, Suicide, and Homicide, 1977-1979 ................. 3 Figure 2. Average annual Age-Adjusted Death Rates for Homicide for Persons under 45 Years of Age, 1979-1981 ...... 7 Figure 3. Homicide Rates, by Color and Sex: United States, 1915-76 .........................11 Figure 4. Death Rates from Homicide by Race and Place of Residence 1977-1979 ................... 14 Figure 5. 5-Year Homicide Rates for the Period 1976-1980 by Ethnicity, Age Group and Sex..............17 Figure 6. Age Adjusted Homicide Death Rates for Native Americans........................19 Figure 7. Death Rates from Unintentional Injury by Age and Race, 1977-1979.....................20 Figure 8. Age Adjusted Accident Death Rates for Native Americans........................21 xv TABLES AND FIGURES (continued) Figure 9. Age Adjusted Suicide Rates for Native Americans ..... 23 Figure 10. Death Rates from Suicide by Age and Race........24 Figure 11. Health Factors Relating to Homicide ........... 30 Figure 12. A Continuum for Preventability ............. 42 xvi REPORT OF THE SUBCOMMITTEE ON HOMICIDE, SUICIDE, AND UNINTENTIONAL INJURIES CHAPTER I INTRODUCTION The Subcommittee on Homicide, Suicide, and Unintentional Injuries is one of six subcommittees formed by the Secretary's Task Force on Black and Minority Health to investigate the major differences in health status of Blacks, Hispanics, Native Americans, and Asian/Pacific Islanders compared to the nonminority population and to establish a framework for the Department of Health and Human Services to improve the health of minority Americans. The statistical technique of "excess deaths" was the primary method employed by the Task Force to measure the health status differences between minorities and nonminorities (1). This method quantified the number of deaths that would not have occurred had mortality rates from intentional or unintentional injuries among minorities equalled those of nonminorities. For example, in Blacks, 40 to 47 percent of the total annual deaths from all causes, were calculated to be excess deaths. Homicide and unintentional injuries are major contributors to excess deaths among Blacks, accounting for 35 percent of the excess deaths in Blacks under age 45, and 18.5 percent of excess deaths in Blacks under age 70. Homicide and unintentional injuries persist as important contributors to excess deaths among Hispanics and Native Americans. Suicide among certain minority groups also was included in the investigations of the subcommittee. This Subcommittee was charged with: • Examining the public health implications of homicide, suicide, and unintentional injuries (accidents). • Identifying major disparities between minorities and nonminorities in the areas of death from homicides, suicides, and unintentional injuries. • Reviewing factors associated with high rates of homicide and other violence in the United States. • Suggesting preventive intervention strategies appropriate for implementation by the public health community and by others with particular attention to the needs of minority Americans. 1 A review of data compiled and analyzed by the Task Force or subsequently obtained by the Subcommittee, led to the identification of seven major areas of excess, unnecessary deaths from violent causes in minority populations. These areas, studied further by the Subcommittee are: • Black male homicide • Black female homicide • Hispanic male homicide • Native American homicide • Native American deaths due to unintentional injury • Native American suicide • Chinese female suicide Figure 1 compares death rates from unintentional injuries, suicide, and homicide among White, Black, Native American, and Asian populations (2). Homicide deaths among Blacks and Native Americans significantly exceed those of the White majority. Deaths from unintentional injuries among Native Americans occur at more than twice the rate for other groups. Although the overall suicide rate among Native Americans does not exceed that of the White population, individual tribes have rates of suicide that are several times greater. Death rates among Asians are below those of the general population for both intentional and unintentional injuries except in the case of Chinese female suicides. Not shown in Figure 1 are the excess homicide deaths among Hispanic males compared to other White males. Evidence of this disparity is presented later in this report. Prevention of homicide, and of Black and minority homicide in particular, is a new endeavor for the health field. Like many other modern health problems, homicide has multiple causes and is not susceptible to simple solutions. After its study of homicide, however, this Subcom- mittee is convinced that important new steps can be taken towards more effective prevention of homicide through application of traditional public health concepts of prevention. Much of the remainder of this report will be concerned with ways in which these concepts of prevention can be applied to the homicide problem as it affects the Nation's Black, Hispanic, and Native American populations. High rates of suicide and unintentional injuries in the Native American population have long been problems of concern to the health sector. In 2 Figure 1 Death Rates from Unintentional Injury, Suicide, and Homicide by Race, 1977-1979 Deaths per 100,000 Population 100 i------------------.----- Unintentional Suicide Homicide Injury SOURCE: The Injury Fact Book, Baker, S., O'NeiU, B., Karrf, R.S. this report, the Subcommittee reviews these problems once more and recommends some new directions for preventive interventions. There is a great dearth of published health data on Asian/Pacific Islanders. The Task Force did not become aware of high suicide rates among Chinese American women until a special study was made of Chinese, Japanese, Filipino, and Native American death certificates submitted by the 50 States to the National Center for Health Statistics. REFERENCES: Chapter 1 1. Department of Health and Human Services, Report of the Secretary's Task Force on Black and Minority Health. Volume 1: Executive Summary Washington, D.C., U.S. Government Printing Office, 1985. 2. Baker, S.P., O'Neill, B., Karpf, R.S.: The Injury Fact Book. Lexington, Lexington Books, D.C. Heath and Company, 1984. 4 CHAPTER II MAGNITUDE OF THE PROBLEM In 1983, homicides accounted for more than 19,000 deaths per year in the United States, a rate of 8.2 deaths per 100,000 population (1). The homicide rate in the United States continues to be significantly higher than that of any other industrialized nation (see Table 1). In 1983, homicide was the eleventh leading cause of death in the United States for all ages and races combined. Because homicide victims are mostly young adults, the actual impact of homicide is measured better by potential years of life lost annually, rather than just by numbers of victims. For Americans of ages 1 through 65, homicide accounts for more than 726,000 potential years of life lost annually, an index for which homicide ranks fourth among all causes causes of death (2). Figure 2 illustrates the average annual age-adjusted death rates for homicide among males and females of different racial groups, under age 45 during the period 1979-1981. Data on homicide deaths among Hispanics are not as complete as those available for White and Black populations. The seriousness of the Hispanic male homicide problem is indicated by data from a special study of five Southwestern states where the majority (60 percent) of Hispanics reside. BLACK HOMICIDE No cause of mortality so greatly differentiates black Americans from other Americans as homicide. In 1983, Blacks constituted 11.5 percent of the United States population but accounted for 43 percent of all homicide deaths (3). Black men, women, and children all have rates of death from homicide that are far in excess of the rates for their fellow citizens of the same age and gender. As shown in Table 2, the death rate from homicide among Black men was more than 6 times the rate for White men. Black females have had consistently higher homicide rates than White males and much higher homicide rates than White females. The homicide rate since 1980 for Black women was 3 times that for White women and approximately 30 percent higher than the rate for White men. The death rate from homicide per 100,000 live births was nearly 3 times higher for Black children under 1 year than for white children in the same age groups (4). Homicide is the leading cause of death for Black males between the ages of 15 and 44, and for Black females between the ages of 15 and 24 (see Table 3). Homicide accounts for more excess mortality I 5 Table 1 Homicide Rates for Selected Countries, 1970 to 1978 Rate per 100,000 population Country 1970 1971 1972 1973 1974 1975 1976 1977 197( Austria 1.5 1.4 1.2 1.4 1.5 1.6 1.3 1.4 Belgium 1.1 1.0 1.2 1.1 1.0 .9 .9 Canada 2.0 2.2 2.3 2.4 2.5 2.4 2.6 Denmark 1.0 .6 .9 .7 .6 .7 .5 Finland mg 2.7 3.2 2.7 2.6 3.3 2.8 France .7 .8 .9 a .9 .9 1.0 Germany (F.R.) 1.4 1.3 1.4 1.2 1.2 r.2 1.3 1.2 1.2 Ireland .4 .5 .7 .4 .7 1.0 .3 .9 Italy .0 1.0 1.1 1.2 1.1 1.4 Netherlands #5 .6 .5 .6 .8 .7 .9 .8 Norway .6 .7 .8 .6 .7 .8 .7 Sweden #3 .9 1.1 1.0 1.2 1.1 1.2 1.0 Switzerland mY 1.0 .9 .7 1.0 .9 .9 .7 United Kingdom .9 J9 1.2 1.0 1.0 .9 1.2 United Slates 7.7 9.1 9.4 98 10.2 10.0 9.1 9.4 Source: Curtis, Lynn (Ed.) American Violence and Public Policy. New Haven, Yale University Press, 1985. 6 Figure 2 Average Annual Age-Adjusted Death Rates for Homicide for Persons Under 45 Years of Age, 1979-1981 Deaths per 100,000 Population \JU Males 80 73.4 60 40 — 24.6 20 — 10.9 WmL 80 n White Black Native American Asian/ Pacific Islander Females White Black Native Asian/ American Pacific Islander NOTE: Death rates for Hispanics are not available. Death rates for Native Americans and Asian/f^cific Islanders are probably underestimated due to less frequent reporting of these races on death certificates as compared with the Census. SOURCE: National Center for Health Statistics, Bureau of the Census, and Task Force on Black and Minority Health. TABLE 2 Death rates for homicide and legal intervention: 1970 to 1983 1970 1975 1979 1980 1981 1982* 1983* No. of homicides 16,848 21,310 22,550 24,278 23,646 22,358 19,300 Number of deaths per 100,000 populationt All ages, 9.1 10.4 10.2 10.8 10.3 9.7 8.2 all races White Males 7.3 9.3 9.9 10.9 10.3 9.5 Black Males 82.1 79.8 70.1 71.9 69.2 62.3 White females 2.2 2.9 2.9 3.2 3.1 3.1 Black females 15.0 16.1 13.9 13.7 12.9 12.0 Ratio of Blacks to Whites Males 11.2 8.5 7.1 6.5 6.7 6.6 Females 6.8 5.6 4.8 4.3 4.2 3.9 * provisional data t age adjusted Source: National Center for Health Statistics; Health, United States, 1983; Health, United States, 1984; and Advance Report of Final Mortality Statistics, 1980, 1981, 1982. 8 TABLE 3 Black Males and Females: Death Rates for 5 Leading Causes of Death, 1981, Selected age groups BLACK MALES Cerebro- Age Group Homicide Diseases of the Heart Cancer Motor Vehicles vascular Disease 15-24 78.2 6.7 7.0 30.8 1.5 25-34 136.9 29.3 14.1 42.2 7.2 35-44 106.1 129.3 75.8 40.0 29.2 BLACK FEMALES 15-24 16.9 4.2 4.6 25-34 23.2 13.7 17.4 35-44 16.3 56.0 73.7 7.7 1.6 8.2 6.6 7.7 21.0 Source: Health United States, 1984 among Black Americans under age 45 than any other cause of death. Up to age 70, homicide is second only to heart disease in its contribution to excess deaths. The phenomenon of extremely high homicide rates in the non-white population is not new. Since 1914, when national mortality data were tabulated for the first time by cause of death and race, death rates from homicide among non-white males have exceeded those for White males by factors of as much as 12 to 1 (see Figure 3). Homicide rates for nonwhite females have consistently exceeded those for both White males and females. National data on the race of homicide victims during the period 1968-1978 are presented in Table 4. The data show a consistent annual trend of proportionally decreasing nonwhite victimization. Despite the trend, Blacks continue to be greatly overrepresented as homicide victims, given the relative proportions of Blacks and Whites in the national population. The lifetime chance of becoming a homicide victim for White persons in 1980 was 1 in 240; for Blacks and other nonwhites, the chance was 1 in 47 (5). In 1983, the Federal Bureau of Investigation reported that the lifetime chance for Black males of becoming a homicide victim was 1 in 21 whereas for White males the chance was 1 in 131. Similarly, Black females had a 1 in 104 lifetime chance of becoming a homicide victim. The chance for White females was 1 in 369 (3). Most homicides are committed by persons who are of the same race as their victims. While there has been some increase in interracial homicide in the United States since the 1960s, the overwhelming majority of Black homicides involve Blacks killing Blacks (referred to as Black- on-Black homicides). A study examining FBI homicide data for the years 1976-1979 found that killer-victim race was the same in 92 percent of all homicides for which the race of killers and victims could be identified (6). In 1983, 94 percent of Black victims were slain by Black assailants, 88 percent of White victims were slain by White assailants (3). Homicides can be divided into four categories based on the nature of the victim-offender relationship: • Family homicide, in which fatal injuries are inflicted by one member of a family on another. • Acquaintance homicide, in which victim and offender are known to each other but are not family. • Stranger homicide, in which the victim and offender have no prior knowledge of each other. • Homicides for which the victim-offender relationship is unknown. 10 Figure 3 Homicide rates by Color and Sex, 1915-1976 80 70 60 Z 2 s, < 2 § 40 8 30 20 10 t \ v A' 1 ii 11 i » i • i % i i i i i i i $ i t i i i i i h K VV 1/ ALL OTHER MALE V^ < ,/~%mm*-**v*\, ALL OTHER FEMALE „*%, ^""^^ X^ -WHITE MALE ^.^# WHITE FEMALE i«>» .no tr» mo ins i*o i*m »no ins two iw '•»« it» « Source: Klebba, A. Joan: Comparison of trends for suicide and homicide in the United States, 1900-1976. In: Violence and the Violent Individual. Hays, Roberts, and Solway (Eds.) New York, SP Medical and Scientific Books, 1981. 11 Table 4 Race of Homicide Victims in the Nation, 1968-1978 Race of Victim White Black Other Ra ce Race Unknown Total Vict tns Year Frequency X Frequency X Frequency X Frequency X Frequency X 1968 5,449 45.6 6,351 53.1 120 1.0 35 0.3 11,955 100.0 196^ 5,740 44.4 6,984 54.1 158 1.2 36 0.3 12,918 100.0 1970 5,812 44.6 7,065 54.2 130 1.0 32 0.3 13,039 100.1 1971 6,840 44.6 8,238 53.8 207 1.4 37 0.2 15,322 100.0 1972 7,158 45.2 8,422 53.2 238 1.5 14 0.1 15,832 100.0 1973 8,031 46.9 8k863 51.8 212 1.2 17 - 0.1 17,123 100.0 1974 9,034 48.5 9,266 49.7 306 1.6 26 0.1 18,632 99.9 1975 9,463 50.8 8,831 47.4 298 1.6 50- 0.3 18,642 100.1 1976 8,475 51.0 7,732 46.6 345 2.1 56 0.3 16,608 100.0 1977 9,470 52.5 8,176 45.3 358 2.0 30 0.2 18,034 100.0 1978 10,111 54.0 8,201 43.8 352 1.9 51 0.3 18,715 100.0 Source: FBI Uniform Crime Reports, Supplementary Homicide Reports, 1968-1978. Notes: Percentages may not sum to 100.0 due to rounding. "Other Race" includes American Indians or Alaska Natives, Asian/Pacific Islanders and persons of Hispanic origin. Black homicides tend to involve acquaintances more often than family members or strangers. A study of 1978 national data found that 53 percent of all acquaintance homicide victims were Black. Persons involved in acquaintance homicides as offenders and victims are usually in their twenties. Acquaintance homicides most often occur within a private residence; one third occur in the street (7). Homicide rates for Black Americans are higher than those for White Americans in both rural and urban sectors of the country. Figure 4 illustrates death rates from homicide by race and place of residence. Rates are lowest in rural areas and in cities of less than 250,000 pop- ulation, higher in larger cities, and reach their peak in cities of 1,000,000 and more. In 1980, about 60 percent of the nation's Black population lived in central cities. Table 5 presents data on homicide rates in the 14 cities with the largest Black populations. With the exception of Memphis, Black homicide rates in all of these cities were in excess of the national Black rate. In three cities—St. Louis, Los Angeles, and Cleveland—the Black homicide rate was more than double the national Black rate. Firearms are the most common means of committing homicides. FBI national data for the years 1971-1983 indicate that about 60 percent of all homicides are committed with firearms: handguns, rifles, or shotguns. National data for the years 1977-1979 indicate that Black homicide deaths are accomplished by firearms at a higher rate than White homicides (8). A study of all deaths of California residents found that firearms were involved in 65 percent of all Black homicides during the five-year period 1978-1982. The same study also showed that firearms-related deaths from all causes (homicides, suicides, accidents) were the leading cause of death for Black male Californians ages 1-54 and that, if the death rate remained unchanged, one in 20 of the Black males born in California during 1978-1982 would eventually have a firearms related death (9). Knives and other sharp instruments are the second most common weapons used in homicide, accounting for 18 percent of deaths. Strangulation, beatings, and falls from high places account for most of the remaining homicides. Death by strangulation causes 12 percent of homicides among females but only 2 percent among males. HISPANIC HOMICIDE More than 60 percent of all Hispanics in the United States reside in the five Southwestern States of Arizona, California, Colorado, New Mexico and Texas. Since 1976, health officials in these five States have noted on death certificates whether the deceased was Black, non-Hispanic White, or Hispanic. In 1984, a research team from the Centers for Disease 13 Figure 4 Death Rates from Homicide by Race and Place of Residence, 1977-1979 50-f 40- c o 2s "5 a o Q. O o o o o 0 White Black 6 Largest Cities Other Large Cities Source: Baker, et al. (8) 14 TABLE 5 Homicide Rates in 14 Major Cities of the United States, 1980 Black Homicide Aggregate Homicide City Rate per 100,000 Rate per 100,000 Blacks ___ ______Population Atlanta 46.3 38.6 Baltimore 40.6 28.6 Chicago 45.1 28.6 Cleveland 76.8 45.7 Dallas 55.3 33.6 Detroit 59.3 45.7 Houston 45.9 37.1 Los Angeles 87.3 33.8 Memphis 37.7 24.2 New Orleans 56.8 37.5 New York 46.8 23.9 Philadelphia 51.8 26.4 St. Louis 91.6 48.8 Washington, D.C. 39.9 36.1 Mean black homicide rate for 14 cities 55 National black homicide rate, approximately 38 Source: Dr. Harold M. Rose. Figures derived from FBI monthly homicide reports for 1980. 15 I Control carried out a special study of death certificates in these five States comparing patterns of White and Hispanic homicide (10). The study covered the years 1976 through 1980. The results of the study indicated that the rates for White and Hispanic homicide victims during the five- year period were sharply different. The overall age-adjusted Hispanic homicide rate was 21.6 per 100,000, more than two-and-one-half times the rate of 7.7 per 100,000 for the White population in the same geo- graphical area. By comparison, the national homicide rate in 1979 was 10.2 per 100,000. Figure 5 illustrates the homicide rates in the Southwestern States for Hispanic and White males and females by age group. In comparing homicide rates between sexes, Hispanic and White females had approximately the same overall homicide rate of 4.8 and 4.2 respectively. The age- adjusted homicide rate among Hispanic men was 39.3 compared to 11.4 for White men, more than a three-fold difference. Furthermore, Hispanic males had a higher homicide rate than White males for every age group studied. The most striking differences occurred among males 20 to 24 years old, the age span in which most homicides occurred for both White and Hispanics. White males of that age had a homicide rate of 18.5 per 100,000 population while Hispanic males of ages 20 to 24 had a rate of 83.3. The finding of an overall Hispanic homicide rate of two-and-one- half times greater than the rate for Whites agrees with two similar studies of cities in the Southwest. Pokorny found that the Hispanic homicide rate in Houston was 2.3 times that of Anglos during the period 1958 to 1961. Loya et al. found Hispanic homicide rates to be 2.2 times that of White homicide rates in Los Angeles for the years 1970 to 1979 (10). The ratios of male to female homicides among Whites and Hispanics is very different in the Southwest. The risk of a White male being a homicide victim is three times greater than the risk for a White female. The risk of a Hispanic male being a homicide victim is five to ten times greater than for a Hispanic female, depending on age. Firearms were the weapon used in 64.1 percent of all Hispanic homicides and in 59.4 percent of all White homicides. Knives were the next common means, but a distant second; 24.8 percent of Hispanic male homicide victims were cut or stabbed to death, while only 17.8 percent of White male victims were killed that way. NATIVE AMERICANS For the purposes of this report, the term Native Americans includes American Indians and Alaska Natives (Aleuts and Eskimos). Data on Native Hawaiians were not included. 16 Figure 5 5-Year Homicide Rates for the Period 1976-1980 by Ethnicity, Age Group and Sex for the 5 Southwestern States ANGLO MALE ANGLO FEMALE HISPANIC MALE HISPANIC FEMALE HOMICIDE RATE PER 100,000 POPULATION 100 AGE Source: Smith, et al. (10) 17 In 1980, the United States Census reported slightly more than 1.4 million American Indians and Alaska Natives residing in the United States both on and off reservations. The data reported here pertain to the approximately 888,000 American Indians and Alaska Natives residing in the 28 reservation states (11) and eligible for health care services by the Indian Health Service (IHS). All data comparisons reported by IHS are made between Native Americans and the general United States population including all races and is designated "U.S. All Races" (12). Considerable variations in culture and tradition exist among the different Native American tribes. These variations, in turn, affect patterns of coping with stress and environment. Although reported as an overall rate, homicide, suicide, and unintentional injury rates among Native Americans vary greatly among individual tribes. Overall, death rates from homicides, suicides, unintentional injuries, and alcohol-related conditions are higher among Native American than among the United States population as a whole. Because these problems occur most frequently among teenagers and young adult Native Americans, and because the Native American population is relatively youthful (33 percent of Indians are under 15), the IHS estimates that the Native American risk of death related to unintentional injuries and violent behavior is likely to remain high for the foreseeable future (13). Native American Homicide The homicide rate among Native Americans has continued to decline since 1974 when it stood at 30.1 per 100,000, almost three times that of the general population (12). In 1980, the homicide rate among Native Americans was 18.1 compared to 10.8 for the general population, 70 percent higher (Figure 6). Unpublished data from IHS show the homicide death rate fell to 14.6 in 1982. Few data are available on homicides involving Native Americans in urban locations. A study by Loya et al. on homicide among American Indians in Los Angeles indicated that the crude homicide rate among American Indian males was very similar to that of White males (14). Native American women in Los Angeles had higher homicide rates than Native American men, and higher rates than their White counterparts, but similar to the rate of Hispanic and Black women. More than 80 percent of the Native Americans were killed by acquaintances. Unintentional Injuries Among Native Americans Of all the minority groups, Native Americans have the highest rate of death from unintentional injuries or "accidents," both overall and at age-specific intervals (Figures 7 and 8). Accidents are the second leading 18 Figure 6 Age Adjusted Homicide Death Rates for Native Americans Source: Indian Health Service, Chart Book, 1984 19 Figure 7 Death Rates from Unintentional Injury by Age and Race, 1977-1979 t—•—i—' i ■—I—'—r—•—r—•—i ■ i—«—i 0 10 20 30 40 50 60 70 80 90 Age Source: Baker, et al. (8) 20 Figure 8 Age Adjusted Accident Death Rates (American Indians and Alaska Natives) Age Adjusted Accident Death Rates 250r~ Per 100,000 Population 200K II 108 90 E Indians and Alaska Natives (Reservation States) U.S. Other than White U.S. All Races i i i , i J—JL I I I I I ' ' ' ' ' ' ' ' 1955 '60 '65 70 Calendar Years *75 '80 Source: Indian Health Service Chart Book, 1984 21 cause of death among Native Americans (13). In 1980, the death rate from accidents among Native Americans was 107.3 per 100,000, significantly greater than the U.S. All Races rate of 42.3 (13). Figure 8 shows the age-adjusted accident death rates for Native Americans compared to U.S All Races. The trend in the Native American accident death rate has been generally downward since 1973, but still remains 2.5 times that of U.S. All Races. Motor vehicle accidents account for the majority of accidental deaths among Native Americans. The overall death rate for motor vehicle accidents in 1980 was 61.3 per 100,000 Native Americans compared to 22.9 for U.S. all races, 2.7 times higher. Nonfatal injuries accounted for approximately a quarter million visits to IHS clinical facilities in 1983. More than seven percent of the IHS budget for clinical services was spent on treatment for injuries. Among the Navajo, the largest of the Native American tribes, the nonfatal unintentional injury rate has climbed steadily and now stands at 184 per 1,000 population, four times greater than that for U.S. all races. Most nonfatal injuries occur near the home. Falls and machinery injuries account for most non-motor vehicle unintentional injuries among Native Americans. Many of the accidents, both fatal and nonfatal, have been attributed to alcoholism or other self-destructive behavior (14). An epidemiologic study of intentional and unintentional injuries among the Hopi Indians was undertaken to determine the circumstances under which injuries occurred, to identify high risk groups, and to develop strategies for injury prevention (15). The overall incidence rate for all injuries was 12 per 1,000 persons per year, with the highest incidence rate, 88 per 1,000, in the over-85 age group. The 15 to 29 year age group, (about one quarter of the population) accounted for 46 percent of all injuries. The leading causes of injuries among the Hopi were falls, motor vehicle crashes, self-inflicted injuries, and assaults. Burns and accidental poisonings were a problem among children younger than 5 years old. Injury problems of special note among the Hopi included single vehicle roll-over crashes, falls from pickup trucks, falls from mesas and pueblo roofs, and suicide attempts in jails. Suicide among Native Americans Suicides among Native Americans peaked in 1977 at 26.6 per 100,000 and have been declining since. The overall suicide rate for Native Americans in 1980 was 14.1, compared to the U.S. All Races of 11.4, about 20 percent greater (see Figure 9). (These figures include suicides only by Native 22 Figure 9 Age Adjusted Suicide Death Rates Age Adjusted Suicide Death Rates 30 (— Pv 100.000 Population Indians and Alaska Natives- US. All Races------- 20f- U.S. Other than White I . i * • * • f j________t i » » I I960 '65 70 Calendar Yaws 75 30 Source: Indian Health Service, Chart Book, 1984 23 Figure 10 Death Rates from Suicide by Age and Race SO-, 40- 3 a. o Q. O o o o" o Q. • O 30- 20- 10- A M / \ I \ I \ t \ I White Black Native American Asian V ^ A *. 7:k X.. \. 7/ \ / / / / -~>.. •**••.— —1—' I ■ I ' I ' 1 ' 1 10 20 30 40 50 60 70 Age 80 I 90 Source: Baker, et al. (8) 24 Americans living in the 28 reservation states.) Because of differences in tribal cultures and traditions, most studies of suicide among Native Americans concentrate on single tribal groups. The considerable variation of suicidal behavior found across tribes makes generalizations about Native American suicide patterns difficult. Navajo and Chippewa, for example, have very low rates of suicide, while Shoshone-Bannock and Apache have rates as high as 10 times the national rate. Suicide has become so severe in some tribes, particularly among the young, that it constitutes a major health problem. The age distribution pattern of suicides among Native Americans as a whole differs from that of the general population; among Native Americans the suicide victims are generally younger, ranging from 15 to 39 years of age. The sharpest peak occurs at age 20 to 24 compared to the general population, where suicides most often occur after age 40 (see Figure 10). ASIAN/PACIFIC ISLANDERS Suicide among Asian/Pacific Islanders A study of Asian American/White American mortality differentials from infancy to adulthood has found only one consistent pattern of excess deaths among Asians (16). As shown in Table 6, older Chinese American women have higher suicide rates than their White female counterparts. Whereas White female American suicide rates peak in the 45 to 54 year age group and decline thereafter, Chinese American female suicide rates rise with each successive age group after 45 and reach their peak in the oldest age groups. After age 75, Japanese American female suicide rates exceed those of White American females. Chinese, Japanese, and Filipino male suicide rates are generally lower than those of American males except in the very oldest groups (16). 25 TABLE 6 Whites, Chinese and Japanese: Average Annual Death Rates for Suicide, United States, 1980 Deaths per 100,000 Population White Chinese Jap anese Age Group Male Female Male Female Male Female All ages, crude 20.57 6.43 8.26 8.28 12.57 6.14 Age-adjusted 19.41 6.20 7.93 8.08 11.08 5.00 5-14 years 0.75 0.28 — 0.61 1.69 — 15-24 years 21.91 5.00 8.07 4.65 14.09 4.52 25-34 years 26.99 7.98 8.59 5.72 16.72 7.82 35-44 years 24.27 9.93 8.94 9.09 12.68 6.39 45-54 years 24.55 11.18 10.77 13.89 9.81 8.22 55-64 years 26.52 9.59 9.37 15.52 12.38 7.78 65-74 years 32.41 7.45 25.85 22.61 11.17 2.17 75-84 years 46.18 6.03 21.82 44.32 39.56 15.75 85 years + 53.28 4.92 64.10 49.93 139.76 19.50 Source: Division of Vital Statistics, National Center for Health Statistics, unpublished data calculated by Elena S.H. Yu, Ph.D. and William T. Liu, Ph.D. Note: In calculating age-specific death rates, the numerator consisted of 1979-1981 cumulative number of deaths and the denominator was based on the total enumerated in the 1981 U.S. census 26 REFERENCES: CHAPTER 2 1. National Center for Health Statistics: Health, United States, 1984. DHHS Publication No. (PHS) 85-1232. Public Health Service. Washington. U.S. Government Printing Office, December 1984 2. Potential years of life lost were computed from age 1 to age 65. Data were computed by the Violence Epidemiology Branch, Centers for Disease Control, from public use data tapes provided by the National Center for Health Statistics. 3. Federal Bureau of Investigation: Uniform Crime Reports for the United States, 1983. U.S. Department of Justice. Washington, D.C, U.S. Government Printing Office, 1984. 4. National Center for Health Statistics. Advance Report of Final Mortality Statistics, 1982 and Monthly Vital Statistics Report, 33:9, December 20, 1984. 5. Centers for Disease Control. Homicide—United States. Mortality and Morbidity Weekly Report 31: November 12, 1982. 6. Jason, J., Strauss, L.T., Tyler, C.W., Jr.: A comparison of primary and secondary homicides in the United States. American Journal of Epidemiology 117: 3, 309-319, 1983. 7. Rosenberg, M.L., Gelles, R. J., et al.: Violence: Homicide, Assault, and Suicide. Closing the Gap health policy project. Data from unpublished working document, 1984. 8. Baker, S.P., O'Neill, B., Karpf, R.S.: The Injury Fact Book. Lexington, Lexington Books, D.C. Heath and Company, 1984. 9. Unpublished data 10. Smith, J.C., Mercy, J.A., Rosenberg, M.L.: Comparison of Homicides among Anglos and Hispanics in Five Southwestern States. Centers for Disease Control, Atlanta, Georgia. 11. A State is considered a "Reservation State" if IHS has responsibilities within the state. Mortality rates refer to the total Native American population of a reservation state including those persons who are outside the IHS service population. The 28 Reservation States are: Alaska Arizona California Colorado Florida Idaho Iowa Kansas Louisiana Maine Michigan Minnesota Mississippi Montana Nebraska Nevada New Mexico New York North Carolina North Dakota Oklahoma Oregon Pennsylvania South Dakota Utah Washington Wisconsin Wyoming 27 12. Indian Health Service, Chart Book Series: June 1984. 13. Listening Post 5: No. 1, February 1984. A periodical of the Mental Health Programs, Indian Health Service 14. Loya, F. et al.: Changes in the criminal homicide rate of American Indians for the city of Los Angeles: 1970-1979. Paper commissioned by the Task Force on Black and Minority Health. January, 1985. 15. Frederick, C.J.: Suicide, Homicide, and Alcoholism among American Indians. National Institute of Mental Health, DHEW Publication No. (ADM) 76-42. 16. Simpson, S.G., Reid, R., Baker, S.P., Teret, S.: Injuries among the Hopi Indians, a population-based survey. Journal of the American Medical Association 249: No. 14, 1873-1876, April 8, 1983. 17. Yu, Elena S.H., Chang, C-F., Liu, W.T., Kan, S.H.: Asian-White Mortality Differentials: Are There Excess Deaths? Paper commissioned by the Task Force on Black and Minority Health. November 1984. 28 CHAPTER III FACTORS ASSOCIATED WITH EXCESS DEATHS FROM HOMICIDE, SUICIDE, AND UNINTENTIONAL INJURIES The United States has a far higher rate of homicide than any other developed country in the world. Black Americans, Hispanic Americans, and Native Americans are distinguished by the fact that their high homicide rates are higher than the already high national average. This chapter will offer some perspectives that can assist in understanding these elevated homicide rates and, in so doing, provide a basis for later identification of ways in which the health sector and others might con- tribute to reducing these rates. Homicide in the United States Considered as a public health problem, the high national homicide rate can be related to a variety of factors operating in one or more of the spheres depicted in Figure 11. • Psychological factors • External environment, including objects within it • Lifestyle, or individual and group ways of life • Physiological factors Psychological factors relate to mental processes and behavior. External environment encompasses physical, historical-cultural, social, and economic environments. Characteristic individual and group behaviors constitute lifestyle. Physiological factors relate to age and gender. Psychological factors. Homicide differs from other health problems in that it is the outcome of mental processes that result in conscious efforts to inflict physical harm upon another human being. Many different types of psychological and psychiatric theories have been developed in an efforat to explain why homicide and other interpersonal violence occurs. Despite this diversity, there appears to be broad agreement that persons who commit homicide and other violent crimes fall into a number of modal groups. These include: 1. Normal, adequately socialized people exposed to extremely provocative or frustrating situations or circumstances. In some instances, their violent tendencies are exacerbated by inhibition-lowering drugs, notably alcohol. 2. Persons committed to a violent lifestyle with supporting attitudes and values. This group includes both normal individuals who consider that violence is appropriate in 29 Figure 11 Health Factors Relating to Homicide Source: Diagram adapted from Health, United States, 1980 30 certain circumstances and other individuals who fail to develop adequate inhibitions against violent behavior because of disturbed developmental patterns. 3. Individuals whose inhibitions against violence are impaired by functional or organic pathology. 4. Overcontrolled individuals whose violence stems from excessive, inflexible inhibitions against the expression of normal aggressive behavior. 5. Individuals who are highly instigated toward aggression or anger for a variety of reasons, including, but not limited to frustration, revenge, jealousy, and oppression. 6. Individuals who engage in violence as a means to achieve goals other than injuring the victim, e.g. robbers (2). While personal characteristics are related to participation in violence, actual violent behavior results from interactions of individuals with their environments. Among the external factors that facilitate or impede violent behavior are cultural attitudes towards violence, socioeconomic conditions, and availability of weapons. Environment. Although occupying one of the most healthful physical environments in the world, the United States has long had higher levels of internal violence than other nations of European origin. As was stated in the report of the National Commission on the Causes and Prevention of Violence, "America has always been a relatively violent nation. Considering the tumultuous historical forces that have shaped the United States, it would be astonishing if it were otherwise " (3). Violence is also accepted in American culture to a degree exceeding that of many other nations. In television programs, movies, and printed media, violence is often presented as entertainment to the American public. The United States places fewer restrictions on private possession and use of firearms than any other Western society. Most of the homicides that occur in the nation are committed with firearms (4). Within the United States, as in other Western countries, rates of homicide and other violent crimes are highest in large cities. Referring to mumerous studies conducted over a period of many years, the National Commission on the Causes and Prevention of Violence noted that violent crime, its offenders and its victims, are most often found in urban areas characterized by low income, physical deterioration, welfare dependency, racial and ethnic concentrations, broken homes, working mothers, low levels of education and vocational skills, high unemploy- ment, high proportion of single males, overcrowded and substandard housing, low rates of home ownership or single family dwellings, mixed land use, and high population density (5). 31 Lifestyle. Several types of behaviors are associated with increased risk for homicide. The National Institute on Alcohol Abuse and Alcoholism has estimated that about one-half of all homicides in the United States are related to use of alcohol (6). An estimated 10 percent of homicides nationwide are associated with use of illegal drugs (7). In some of the nation's largest cities, the number of drug-related homicides is more than 20 percent of all homicides (8). Many, if not most homicides are preceded by patterns of nonfatal violence that can provide targets for efforts at prevention (9). Other research indicates that homicide often is the end product of youthful training in the use of violence that began in the home or through interactions with peers in school and communities (10). Children who engage in antisocial and violent behavior from an early age are at significant risk for persisting in such behavior over time and for becoming deficient in a variety of social skills, including education, communication, and work habits, that are essential to a well-adjusted adult life (11). Physiological factors. Being male and being young are the most important biological risk factors for involvement in homicide as a victim or as a perpetrator. Some homicide offenders are also characterized by organic brain dysfunction or other biological disorders. The impact of biological factors on homicide is mediated by psychological, environ- mental, and lifestyle influences (1). Black Homicide High rates of Black homicide in the United States cannot be explained solely on the basis of factors such as those already cited. Such rates must also be viewed in relation to a Black experience that differs radically from that of any other group in American society. For most of their history in this country, Blacks have been victims, not initiators, of violence. In the slavery and post-slavery old South, violence against Blacks was frequent and supported by law and custom. Violence and crime within southern Black communities often went unpunished since these were matters of little interest to the power structure (12). After 1940, Blacks and their families left the South by the hundreds of thousands to seek work and improved living conditions in northern cities. For lack of education and job skills, most were shunted into deteriorating inner cities where poverty and crime abounded. The term "ghetto" was reinvented to describe city sectors where Blacks were as effectively segregated by racism and skin color as they had been in the South (13). A minority of Blacks ultimately succeeded in making major social gains as opportunities for advancement opened up in education, employment government, and business. The great majority continued to lag far 32 behind White Americans in income, education, employment, and quality of housing. During the 1970s Black scholars and others made increasing use of the term "underclass" to describe large segments of the Black population that now seemed almost permanently trapped in conditions of poverty, joblessness, and welfare dependency (14). Attempts at explaining high Black homicide rates have often focused on the greater poverty of the Black population as the most important factor. Some analyses do, in fact, indicate that when poverty, race, and regional cultural factors are related to Black and White homicide rates, poverty emerges as the most significant correlate of homicide (15). Others have pointed out, however, that Blacks commit much more homicide than Hispanics living under equal or worse poverty conditions in the United States (16). Others have attributed high Black homicide rates to psychological scars inflicted by racism. Particularly among low-income Blacks, this damage is thought to be reflected in feelings of low self-esteem, self-hatred, and rage that are conducive to violence against others (17). Others attribute Black homicide rates to the frustrations engendered in low-income Blacks by life in a society that still discriminates against persons on the basis of skin color. Lacking means to function successfully in the larger social area, this view suggests that low-income Blacks aggressively defend what is left of their integrity in the circle of their families, friends, and acquaintances. This is why so many Black homicides develop out of quarrels over seemingly trivial issues (18). Still another view suggests that the experience of racism and segregation, first in the South and later in the North, has helped to create distinctive "subculture of violence" among low-income Blacks. According to this theory, low-income Blacks are unusually violent because they have evolved a system of values that condones violence in child-rearing and adult interpersonal relationships, sometimes with homicide as a result. Violence is also associated with values that emphasize physical toughness, thrill seeking, sexual prowess, manip- ulation of others, and ready access to weapons (19). Subculture of violence theory has been criticized for its inability to explain how or why the alleged subculture emerged, or why some who are exposed to the subculture become violent while others similarly exposed do not (20). Another criticism is that the theory focuses on value orientations of individuals to the neglect of structural conditions in American society that foster high rates of Black interpersonal violence and homicide. Among the latter are widespread poverty, the lesser value which legal and social institutions in America have traditionally placed on black life, and the tendency of law enforcement agencies and others to attach less importance to violence that affects only Blacks (21). I 33 The health sector can do little on its own to alter the conditions of poverty that adversely affect so many millions of Black Americans. More can be done to affirm the value and dignity of Black life. By iden- tifying reduction of homicide deaths among young Black males as one of its 1990 objectives, the nation's health community made clear its deter- mination to prevent wastage of Black life by homicide insofar as its means would allow. By establishing this Task Force, new impetus has been given to the continuing search for creative solutions. HISPANIC HOMICIDE There has been little systematic study of Hispanic homicide in the mainland United States. National data on Hispanic crime and violence were not collected before 1980 even though Hispanic groups had long urged routine reporting of such data (22). Most attempts to explain Hispanic homicide rates still depend upon partial data collected in various locations at different points of time. Comparative studies of homicide rates in the major Hispanic population groups—Mexican Americans, Central and South Americans, and Puerto Ricans—are not yet available. Rates of Hispanic female homicide appear to be low in most parts of the country. Analyses of data collected on such homicide in Los Angeles and Chicago suggest the existence of strong cultural sanctions within the Hispanic community against male use of lethal force against women. These sanctions also appear to reduce the need of Hispanic women to kill in self defense (23). High rates of Hispanic male homicide appear attributable to the generally low socioeconomic status of the Hispanic male population in the United States, the tendency in all societies for higher rates of homicide to be associated with low socioeconomic status, the macho tradition in Hispanic culture, and the fatalistic expectation of violence and death also associated with this culture (24). Hispanics in the United States have also become highly urbanized: 83 percent lived in metropolitan areas as of 1980 compared to 77 percent of non-Hispanic Blacks and 66 percent of Whites. Much of the urban Hispanic population is clustered in major cities where homicide rates are higher than elsewhere in the United States. More than two million Hispanics now live in Los Angeles; about 1.5 million live in New York City (25). In 1981, Blacks and Hispanics in New York City had identical arrest rates for homicide that were four times greater than the White rate (26). Gang violence is sometimes an important contributor to Hispanic male homicide deaths that occur in the nation's largest cities. In Los Angeles Hispanic homicides classified as "gang-related" accounted for 13.2 percent of all Hispanic homicides during the period 1970-1979 (27). 34 In general, the available data on male homicide indicate that the Hispanic rate is typically less than the Black rate and more than the White rate when socioeconomic level is held constant. NATIVE AMERICANS The Indian Health Service (IHS), a component of the U.S. Public Health service, has been responsible for providing federally-sponsored comprehensive health services to the American Indians and Alaska Natives since 1955. The mental health program of the IHS, established in 1966, recognized the stresses on the lives of many American Indians and Alaska Natives caused by the effects of poverty, forced abandonment of traditional ways of life, inadequate schooling, disruption of the family, and a harsh physical environment. These elements frustrated many American Indian and Alaska Native people in their attempt to live self-respecting and productive lives. The results of these conditions were associated with a high incidence of violence including unintentional injuries, suicides, homicides, and child neglect. Furthermore, each of these elements highly correlated with the concomitant use or abuse of alcohol. Among the various racial/ethnic groups addressed in this report, Americans Indians have the highest frequency of problems associated with alcohol use. American Indian adolescents have significantly higher proportions of heavy drinkers than other groups of adolescents. Estimated rates of alcoholism among American Indians are twice the national average. American Indians have a high arrest rate for crimes related to excessive drinking. Suicide among Native Americans The overall suicide rate for Native Americans in 1980 was 20 percent greater than that of the general population; this differential, however, was the smallest in 20 years (13). The age distribution pattern of suicides among Native Americans is different from that of the general population. The period of greatest risk for Native Americans is the 15 to 24 year age group. The suicide rate for the United States as a whole tends to rise with age. Although reasons for suicide among Native Americans may differ according to tribal and cultural traditions, some generalizations about the contributing factors may be made. Most suicide victims are unmarried males under 30 years of age. According to IHS estimates, alcohol is involved in 65 to 80 percent of suicides (16). Several factors contributing to contemporary Native American suicide are similar to those responsible for alcohol abuse. They include i 35 e Cultural conflict. Native Americans of virtually all tribes have difficulty in relating to the great cultural differences between the dominant American culture and their tribal traditions. The younger Native Americans are caught between old values and customs and their desire to absorb the ways of the White world. They grow up without a satisfactory identification either with their own heritage or that of White society- Native Americans who live in urban areas find adjustment difficult in terms of social and economic survival and value orientation. • Social disorganization. Pressures on Native American society, most notably the forced, rapid social change imposed by the government on American Indians have resulted in a breakdown or disorganization of traditional American Indian socio-cultural systems. Boarding schools run by the Bureau of Indian Affairs also helped to break down cultural systems by attempting to force assimilation on young American Indians (16). • Unemployment or lack of meaningful work. The unemployment rate on some reservations is as high as 80 percent. When American Indians leave the reservation, they have few employable skills. Additional factors that place the young American Indian at high risk for suicide include family disorganization, economic insecurity, family histories of violent deaths, irrelevant schooling, prejudice, crime, and alcoholism. Homicide among Native Americans Although homicide is a serious problem among Native Americans, the rate is considerably less than that for other non-Whites. Suicide is a much greater problem for Native Americans than homicide, especially among the younger age groups. Individuals may be more likely to turn their aggression inward and be more self-destructive than destructive to others. Homicide is more likely to occur when the person is under the influence of alcohol. Among Native Americans, homicide is rarely planned but usually occurs spontaneously as a result of drinking brawls or arguments at home. No analysis of the recent decline in the Native American homicide rate is available. Homicidal behavior among Native Americans is a relatively unexplored area, requiring further research. 36 ASIAN/PACIFIC ISLANDERS Asian American Female Suicide Research is greatly needed to assist understanding and prevention of the high suicide rates among Chinese American women over the age of 45. One significant factor influencing suicide may be the strong stigma attached to mental health problems in the Chinese communities. This can prevent Chinese women who are suffering from anxiety, depression, and the "displaced homemaker syndrome" from seeking needed care, especially if the women fear being "shamed" in the close knit ethnic communities to which they are likely to belong (28). 37 REFERENCES: CHAPTER III 1. Mednick, S.A., Pollock, V., Volavka, J., Gabrielli, W.F. Biology and violence. In: M.E. Wolfgang and N.A. Weiner, (Eds.) Criminal Violence. Beverly Hills: Sage Publications, 1982. 2. Megargee, E.I. Psychological correlates and determinants of criminal violence. In: M.E. Wolfgang and N.A. Weiner, (Eds.) Criminal Violence. Beverly Hills: Sage Publications, 1982. 3. National Commission on the Causes and Prevention of Violence: To Establish Justice, to Ensure Domestic Tranquility. Washington, D.C, U.S. Government Printing Office, 1969. 4. Federal Bureau of Investigation, Uniform Crime Reports, 1983. 5. National Commission on the Causes and Prevention of Violence. 6. John, H.W., Alcoholism and criminal homicide: A review. Alcohol Health and Research World 2: 8-13, 1978. 7. Estimate developed by Research Triangle Institute, North Carolina. 8. New York City Police Department: Homicide Analysis, 1982. 9. Hawkins, Darnell E.: Longitudinal-situational approaches to under- standing Black-on-Black homicide. Paper commissioned by the Task Force on Black and Minority Health, December 1984. 10. Petersilia, Joan: Criminal career research: A review of recent evidence. In: N. Morris and M. Tonry (Eds.) Crime and Justice, Volume 2. Chicago, University of Chicago Press, 1980. Fagan, J.A. et al.: Violent men or violent husbands? Background factors and situational correlates. In: D. Finkelhor et al. (Eds.) The Dark Side of Families. Beverly Hills, Sage Publications, 1983. 11. Loeber, R.: The stability of antisocial and delinquent child behavior: A review. Child Development 53: 1431-1446, 1982. Hartstone, E., Hansen, K.V.: The violent juvenile offender: An empirical portrait. In: R.A. Mathias et al. (Eds.) Violent Juvenile Offenders: An Anthology. San Francisco, National Council on Crime and Delinquency, 1984. 12. Myrdal, G.: An American Dilemma. New York, Harper & Row, 1944. 38 13. Clark, K.B.: Dark Ghetto: Dilemmas of Social Power. New York, Harper & Row, 1965. 14. Glasgow, D.G.: The Black Underclass: Poverty, Unemployment, and Entrapment of Ghetto Youth. San Francisco, Jossey-Bass, 1980. 15. Jason, J., Strauss, L.T., Tyler, C.W., Jr.: A comparison of primary and secondary homicides in the United States. American Journal of Epidemiology 117: 3, 309-319, 1983. 16. Silberman, C.E.: Criminal Violence, Criminal Justice. New York, Random House, 1978. 17. Poussaint, A. F. : Black-on-black homicide: A psychological-political perspective. Victimology 8: 161-169, 1983. 18. Bulhan, H. A.: Frantz Fanon and the Psychology of Oppression. New York. Plenum Press. In press. 19. Curtis, L.A.: Violence, Race, and Culture. Lexington, Lexington Books, 1975. Wolfgang, Marvin E., Ferracuti, F.: The Subculture of Violence: Towards an Integrated Theory in Criminology. Beverly Hills. Sage Publications, 1982. 20. Wolfgang, Margaret E., Zahn, M.A.: Homicide. Encyclopedia of Crime and Justice. New York. The Free Press, 1984. 21. Hawkins, D.F.: Black and white homicide differentials: Alternatives to an inadequate theory. Criminal Justice and Behavior 10: 407-440, 1983. 22. Crimen y justicia, Crime and justice for Hispanics. Discussion paper #1, National Council of La Raza, 1979. 23. Loya, F. et al.: The relative risks of homicide among Anglo, Hispanic, Black, and Asian victims in Los Angeles, 1970-1979. Paper commissioned by the Task Force on Black and Minority Health, December, 1984. Zimring, F.E., Mukherjee, S.K., Van Winkle, B.: Intimate violence: A study of intersexual homicide in Chicago. The University of Chicago Law Review 50: 910-930, 1983. 24. Wolfgang, Marvin E., Ferracuti, F.: op cit. # 19b Rosenquist, CM., Megargee, E.I.: Delinquency in Three Cultures. Austin, University of Texas Press, 1969. 39 25. Hispanics: Challenges and Opportunities. A Ford Foundation Working Paper. 26. New York City Police Department. Homicide Analysis, New York City, 1981. 27. Vargas, L.A., et al.: Gang Homicides in Los Angeles, 1970-1979. Paper commissioned by the Task Force on Black and Minority Health, December, 1984. 28. Chang, S-H.: Mental health delivery systems to Asian Amercians. In: A.K. Murata and J. Farquhar (Eds.) Issues in Pacific/Asian American Health and Mental Health: Report of a P/AAMHRC Task Force. Chicago, Pacific/Asian American Mental Health Research Center, 1982. 40 CHAPTER IV STRATEGIES FOR PREVENTIVE INTERVENTION Until very recent times, homicide was regarded as a problem for the criminal justice system. Health personnel collected data on homicide as a cause of mortality, but responsibility for confronting the problem rested with the police and the courts. The criminal justice system sought to deter homicide through the apprehension and punishment of offenders, but realized that these activities did not constitute an effective homicide prevention strategy. For many years, the annual editions of the FBI Uniform Crime Reports have routinely stated, "It has long been recognized that murder is primarily a social problem over which law enforcement has little or no control." In the reordering of the national health priorities that resulted in the Surgeon General's 1979 report, Healthy People (1), homicide was identified as a problem for which the health sector needed to assume greater responsibility. Subsequently, in the Public Health Service's report Promoting Health/Preventing Disease (2), reducing homicide deaths among young black males ages 15 to 24 was identified as a national health priority. Except for a recommendation that ways should be found to reduce the number of handguns in private ownership, neither report identified specific action that might be taken by the health sector in an effort to prevent homicide. This chapter seeks to remedy the omission. It does so in the recognition that the health sector currently has little to go on in the homicide area. Figure 12 illustrates a variety of health problems problems arrayed along a continuum of preventability. Homicide is located close to the end of the scale that depicts health problems for which there are as yet no known means of effective prevention short of drastic measures (for example, total handgun confiscation) that would be unacceptable to large sectors of the American public or inaconsistent with the American way of life. Enough is known about homicide risks, however, to suggest some useful starting points for applying public health concepts of primary, secondary, and tertiary prevention. As interventions in these areas are developed and tested over time, some demonstrably effective prevention strategies may emerge. PRIMARY PREVENTION Primary prevention in the public health sense involves averting the initial occurrence of a disease, defect, or injury. In the case of 41 Figure 12 A Continuum for Preventability wA^^^^^^^^y.^;.;.^j^^^^^^^^^^^^^^! S^SS:*: :•:$: i N° *novwn $ ^ww«"i$:$:::$i":$::i f Prevent'on$ Smallpox Measles Poliomyelitis Lung and other cancers of the respiratory system Asbestosis Dental caries Cancer of cervix Congenital anomalies Infant mortality Cardiovascular disease Stroke Trauma from accidents Cancer of bladder Pneumonia and influenza Suicide Homicide Brain tumors Rheumatoid arthritis Source: Health, United States, 1980 42 homicide, primary prevention efforts need to be directed at social, cultural, technological and legal aspects of the environment in the United States that facilitate the perpetuation of the nation's extra- ordinarily high homicide rates. The following examples illustrate the types of preventive strategies that need to be developed. Implementation of these strategies will require that health professionals join with others in an effort to eradicate factors that impair health by facilitating homicide. Public Education: A national health promotion campaign to prevent homicide Effective health strategies to prevent homicide must begin by en- listing greater public and professional interest and concern. Although survey data are lacking, there appears to be little general awareness that homicide is a health problem as well as a criminal justice problem. Nor does it seem widely understood that homicide is one of the nation's major causes of premature death, that homicide rates in this nation far exceed those of any other industrialized nation, and that prevention and reduction of homicide should be achievable if more concerted attacks are made on this serious health problem. The Public Health Service should take the lead in developing a new health promotion campaign specifically focused on homicide. The goal of the campaign should be to increase public and professional awareness that homicide is a serious national health problem and that it is preventable. By pointing to American homicide rates as compared to those of other nations, the campaign can also convey the message that current high rates of American homicide are intolerable. Particular attention should be given in the health campaign to effects of homicide on Black, Hispanic and Native Americans. It seems doubtful that Black Americans sufficiently realize that homicide is the largest single killer of young Black men and women and that these deaths are preventable. It is doubtful, also, that the Hispanic and Native American communities are aware of the gravity of their homicide problems. Urban data on homicide needs to be incorporated into health promotion materials because of the high rates of Black, Hispanic, and other homicide in many of the nation's larger cities. To the extent possible, the data on homicide should also be linked to data on nonfatal violence to emphasize that health strategies aimed at homicide must also be concerned with non- fatal types of violence that often precede and are contributory to homicide (e.g. child abuse, spouse abuse, rape, robbery, and assault). Education of Health Care Providers Physicians and other health care providers need to develop greater understanding of homicide as a public health problem for which the 43 health sector needs to accept greater responsibility. Information on homicide and other violence needs to be incorporated into the curricula of medical schools, nursing schools, schools of social work, and continuing professional education. Community Self-Help Traditionally, communities characterized by high homicide rates have tended to believe that little could be done to reduce these rates through means other than law enforcement and the application of criminal sanctions. This view is now giving way to another. Increasingly in Black communities, high rates of Black-on-Black violence and homicide are being regarded as a problem for which Blacks themselves can accept greater responsibility. Similar trends can be encouraged in Hispanic, Native American, and other communities. "Community ownership" of a homicide problem has important implications for efforts at prevention. Homicides are not events mandated by some biological or social necessity; in theory at least, every homicide is avoidable and preventable. By accepting responsibility for homicides that occur within their own communities, citizens empower themselves to press more actively for adequate and comprehensive efforts at prevention. Community-based citizen crime prevention is an emerging idea that also is well supported politically. Fattah (3) and Heifer (4) suggest developing concurrent community approaches that might include some combination of the following: 1. A community consortium or community council composed of civic, political, religious, youth, and other community leaders, committed to the proposition that violence and high homicide rates in the community are unacceptable and preventable. This group would meet regularly to exchange ideas and information and, with help from other experts, if needed, would develop constructive approaches to prevention of violence and homicide. 2. A continuing media campaign to educate the community and the general public that violence and homicide are unacceptable. 3. Development of an information bank on violence and homicides that occur in the community, with special attention to the role of alcohol, illegal drugs, and firearms in such violence. 4. Examination and monitoring of community resources, including economic, educational, social, cultural, health, and law enforce- ment to maximize potentials for violence reduction and homicide prevention. 44 5. Development of a community policy towards alcohol sales and drug abuse. 6. Development of a community policy toward firearms. In a free society, Black and minority communities have the opportunity to consider the firearms problem and chose a policy that is in accordance with their own community needs. The policy choices do not have to be the extremes of a firearms ban or no firearms restrictions whatsoever. Moore (5) has proposed an alternative approach that community leaders may wish to consider: a. Recognize that firearms have legitimate uses in American society. b. Focus on keeping handguns away from criminal offenders and off city streets instead of trying to disarm the entire community. Keeping guns from criminal offenders also depends on controlling household thefts of handguns and local handgun black markets, tasks for which the local police departments are well suited. c. Keep guns off city streets through legislative and law enforcement means. The idea that carrying a gun in congested urban areas is a dangerous activity that cannot be justified seems to have widespread social support that is reflected in numerous "carrying" statutes and ordinances around the nation. Interventions against televised violence Many researchers have concluded that televised violence has a causal effect on aggressive or violent behavior among children and youth. Television networks and other supporters, on the other hand, contend that television programming reflects the violent behavior already present in society. Government and television network controls have not been sufficiently effective in reducing televised violence. Government may be reluctant to apply restrictions on television violence because of concerns about censorship. Private actions, therefore, may be more acceptable than government action. Health education efforts should be planned and implemented to make parents aware of the dangers of televised violence. Program selection guidelines developed by consumer groups should be publicized so parents may choose programs that demonstrate appropriate prosocial behavior. Several consumer and professional groups have already organized against violence on television, for example, the American Medical Association, the National Parent Teachers Association, and a consumer group, Action for 45 Children's Television have been actively campaigning for modifications in television programming. Further studies are needed to evaluate the effect of television as a potential risk factor in promoting violence, cigarette smoking, alcohol use and risky driving. The potential role of television to support health promotion through positive role models should also be investigated. Active industry participation should be sought in any effort relating to television programming. SECONDARY PREVENTION Secondary prevention in the public health sense involves halting or slowing the progression of a health problem. Early detection and case finding are the means by which future, more serious morbidity may be decreased. In the case of homicide, such case finding requires identification of individuals manifesting early signs of behavioral and social problems that are logically and empirically related to increased risks for subsequent homicide (6). Many, if not most homicides are preceded by patterns of nonfatal violence that can provide targets for efforts at secondary prevention (7). Research indicates that homicide often is the end product of youthful training in the use of violence that began in the home or through interactions with peers in school and communities (8). Children who engage in antisocial and violent behavior from an early age are at significant risk for persisting in such behavior over time and for becoming deficient in a variety of social skills, including education, communication, and work habits, that are essential to a well-adjusted adult life (9). Family violence, childhood aggression, school truancy and drop out, school violence, and adolescent violence are important focal points for efforts at secondary prevention of homicide. In the case of Black and Hispanic homicide, such preventive efforts are particularly needed in low-income, inner city communities characterized by high rates of violence, family disorganization, unemployment, and school drop out. Family Violence Each year in the United States, an estimated 2.9 million American households are the scene of severe husband-wife violence. An estimated 6.5 million children are the victims of severe violence inflicted by their parents (10). Although a great many children who grow up in violent homes do not become violent themselves, exposure to severe and repeated violence in one's childhood home appears to be an important precursor of subsequent violence as a child and as an adult (11). 46 Few studies thus far have attempted to examine racial and cultural differences in family violence. The available data suggest that Black families may have the highest rates of child abuse and neglect, followed by Hispanics and Whites. The data also indicate that Black maltreating families are mostly poor, on public assistance, and headed by a single parent who has not graduated from high school (12). Other research indicates that higher rates of child abuse in low-income Black families may be attributable to a cultural belief that hitting children is a necessary method of childrearing (13). A national survey of family violence conducted in 1976 found that violence against wives was nearly four times more common in Black families than in White families, and that violence against husbands was twice as common (14). No comparable data currently exist for Hispanic families. In 1985, the National Institute of Mental Health (NIMH) funded a new national survey of physical violence in American families. Data was collected from representative national samples of White, Black, and Hispanic families (15). Individuals from single parent families, unmarried cohabitating couples, and recently terminated marriages are included in the samples. Data from this survey can greatly assist in the identification of needed violence prevention efforts. An example of a culturally relevant intervention that can be employed is the Culturally Adapted Training Project supported by NIMH. This project involved adaptation of three standard parent training programs to the needs and life circumstances of low-income Black families. New instructional units were developed on traditional Black discipline, modern Black self-discipline, Black pride, and single parenting. The purpose of the two units on discipline is to assist low-income Black parents in developing more positive, affectionate parenting styles that could reduce and eliminate reliance on hitting as a means of child rearing (16). Child Aggression High levels of aggressive behavior in preadolescent children, partic- ularly when coupled with rejection by well-behaving peers, is a significant risk indicator for antisocial and violent behavior in later life. Aggressive behavior tends to be stable over time, with the result that children who were highly aggressive at one stage in their development are likely to remain so at later points in childhood and adolescence (17). The stability of aggressive behavior over time is greater than that of most other childhood behavior patterns and—particularly when some measure of severity is taken into account—is the single most powerful predictor of violence and other disorders in later life (18). A tendency to take offense too easily or to infer hostile intentions from others has been identified as a primary factor in preadolescent 47 aggressive behavior (19). Such a tendency has been a major source of vulnerability for Black children, especially boys, because of the importance that facility and poise in verbal play has had for the social acceptance and status of Black adolescents (20). Several promising models exist at the present time for preventive interventions with aggressive children that could reduce subsequent risks for violence and homicide. These include two models that have had promising results when used with low-income Black preadolescents. The first addresses the problem-solving deficits that are often observable in aggressive children and provides systematic, step-by-step remedial training (21). The second is an Anger Coping Program that has been developed for use with aggressive elementary school boys and is designed to reduce impulsiveness and promote problem-solving (22). School Truancy and Drop-out The importance of staying in school and completing high school has been amply documented in numerous studies. In addition to reducing chances for economic well-being in later life significantly, school truancy and drop-out in the adolescent years are linked to increased risks for involvement in crime and violence. • A study of a predominantly Black school district in North Carolina found that of the junior high school students who were reported truant in one school year, 90 percent subsequently dropped out of school and 40 percent were arrested during the following three-year period (23). • A study of Black males in St. Louis found that serious truancy (absence of more than 20 percent of the time in five school quarters) was the strongest predictor of early death from homicide and other causes (24). Preventing truancy and drop-out are important steps that could be taken to reduce subsequent, more serious problems including homicide. Hispanics are especially disadvantaged in this area and have drop-out rates that are more than double the national average (25). School administrations traditionally have used suspension from school as a means of dealing with chronic truancy and other serious student misbehavior. Data collected during the 1970s showed that Black students were receiving a highly disproportionate share of suspensions from school, that the affected students were often those who could least afford to miss academic instruction, and that suspension from school was associated with premature school drop-out. Many schools are now making increased use of in-school alternatives in order to avoid the harmful effects of out-of- school suspension (26). 48 School Violence Public junior and senior high schools serving low-income populations in large cities can be dangerous environments. • A 1975-1976 national survey found that 40 percent of the robberies and 36 percent of assaults that victimize urban youth occur in high schools (27). • A 1983 survey of four high schools in Boston found that 37 percent of the male students and 17 percent of the female students reported carrying a weapon to school at some time during the school year. Half of the public school teachers in Boston had been victimized by crime on school property at least once during the same school year (28). With strong encouragement from President Reagan, many states and school districts have formed task forces to address problems of school discipline, violence, and crime (29). In 1983, the Department of Justice in cooperation with the Department of Education, awarded a grant to permit formation of a National School Safety Center to assist schools and communities in addressing the whole spectrum of school safety issues (30). Undergoing these efforts has been the recognition that students learn from the behavior of others, and if violence and crime occur in schools and go unpunished, a climate is created that can encourage other violence (31). Public school health education courses are being used as a strategy for preventing interpersonal violence and its extreme, homicide. Fights among Black students in inner city schools often have characteristics similar to much Black-on-Black homicide-the fights are usually b^ween students who know each other and who get into arguments in which drugs and alcohol may have significant roles. The Boston Youth Program based at Boston City hospital and supported by the Robert Wood Johnson Foundation, has developed an innovative curriculum which has been ^^f/^0 mandatory tenth grade health classes. The curriculum, geared to Black students, is aimed at assisting students to cope more effectively with problems of anger, fighting, and violence in their daily life (32). Another approach to crime and violence in high schools has been the introduction or law-related curricula. Florida has developed a Program in cooperation with the state bar association that teaches students the onsequencL of breaking the law and covers a wide range o ^districT issues. As of 1985, approximately 70 percent of florida school districts were using the program (33). Chemical Dependency Programs to reduce chemical dependency need to be incorporated into Strategies aimed at prevention of homicide, suicide, and unintentional injuries in minority populations. A national survey of prison inmates 49 found that 50 percent of convicted Black homicide offenders reported drinking at the time of the crime (34). A study of all homicide deaths in Los Angeles during the period 1970-1979 found that alcohol was detected in 57 percent of Hispanic victims and 47.7 percent of Black victims (35). It has been estimated that 75 to 80 percent of Native American suicides are alcohol related, and that alcohol is heavily involved in Native American deaths from unintentional injuries and homicide (36). Few data are available on Black and minority deaths related to illegal drugs. A study of 546 predominantly Black and Hispanic homicide victims in New York City found that 46 percent could be classified as substance abusers (37). In 1984, homicide detectives in the District of Columbia classified 20 percent of homicides in this heavily Black city as drug related (38). Intervention strategies that can assist in reduction of alcohol abuse and drug abuse in the nation's Black, Hispanic, and Native American populations are discussed in the report of the Subcommittee on Chemical Dependency of the Task Force on Black and Minority Health. TERTIARY PREVENTION Tertiary prevention is concerned with situations in which a health problem is already well established, but efforts can still be made to prevent further progress toward increased disability and death. In relation to homicide, the problems of greatest concern are types of interpersonal conflict and nonfatal violence that appear to have a high risk for homicide. Aggravated Assaults Police agencies usually classify any assault causing physical injuries that requires medical treatment beyond first aid as an aggravated assault. Attempted murders and assaults with a deadly or dangerous weapon are also counted as aggravated assaults (39). Nationally, arrests of men for aggravated assault exceed arrests of women by about 15 to 1. Arrests for aggravated assaults exceed arrests for murders by about 12 to 1 (40). Because data from hospital emergency rooms suggest that four times more cases of aggravated assault are seen in hospitals than appear in police records (41), the true ratio of aggravated assaults to homicides may be closer to 50 to 1. The relationship between aggravated assault and subsequent homicide is suggested by a study of all homicides that occurred during a two year period in Kansas City. In about 25 percent of the homicides, either the victim or the offender had previously been arrested for an assault or a disturbance (42). Assaults that bring victims to hospital emergency rooms 50 are thus an important potential means of identifying individuals at risk for future homicide. Differential rates of Black, Hispanic, and White homicide also appear to be mirrored in assault cases that come to hospitals, • A study of 41 acute care hospitals in Northeastern Ohio (Cleveland and Lorain-Elyria SMSAs) found that the city of Cleveland had an assault rate that was much higher than that of any other jurisdiction. Cleveland is the only jurisdiction in the study area with a large, inner city Black population (43). • A Chicago study found that interpersonal violence was the leading cause of head injury cases in a depressed inner city Black sector of that city, and that the rate of head injury from this cause was far higher than for suburban Whites (44). • A New York City study found that interpersonal violence was the leading cause of head injury among males admitted to three acute care hospitals in an area of the city characterized by extensive poverty and high rates of crime, substance abuse, and unemployment. Rates for head injuries caused by violence were highest for Black males, intermediate for Hispanic males, and lowest for White males. An identical distribution was observed for Black, Hispanic, and White females who had head injuries due to violence (45). More research will be needed before efforts to prevent homicide can incorporate hospital emergency rooms into their designs. Meanwhile, the available data suggest a number of useful steps than can be taken in emergency rooms and hospitals towards this end. • Improve medical record-keeping on assaults. Currently, these records tend to contain only a description of the physical injury. Very little is recorded about the patient, the perpetrator, and the circumstances of the assault. Unlike other disease processes, injuries suffered from assaults are generally recorded and regarded as isolated events even for those persons for whom there is a previous hospital record of assaults. • Alcohol, illegal drugs, and drug transactions are often associated with aggravated assaults and related risks for homicides. Inform- ation on use of these substances needs to be incorporated in the hospital record. Blood alcohol levels should be routinely measured. Although victims of aggravated assault who come to hospital emergency rooms appear to be, as a group, at high risk for homicide, successful preventive intervention with such persons may be difficult. A study of male assault victims admitted to one large urban hospital found that the typical victim was an unemployed or underemployed single Black male in his mid-thirties with limited education, few marketable skills, a tendency to abuse alcohol, and a number of health problems. His involvement in street violence and other occasions for violence was frequent (46). 51 Spousal Violence One of the most striking features of homicides committed by and against women is the intimate setting in which these homicides most typically occur. When they murder, women are much more likely than men to murder family members and other intimates. In contrast to murders that occur among men, murders by women almost always involve a member of the opposite sex. These patterns hold true for Black women even though their murder rates far exceed those of White females (47). Studies conducted at major urban hospitals have shown that battering by a husband or other male intimate is a frequent source of the physical injuries and other complaints that bring women to hospital emergency rooms (48). Research has also indicated that about 20 percent of the battered women who come to emergency rooms are enmeshed in a "battering syndrome" characterized by multiple physical injuries over time and psychosocial disorders that include psychiatric problems (49). Data are not yet available on the relationship of wife battering to family homicide. Strong indications, however, point to spousal violence as a two way street, with women often responding to violence with violence (50). These considerations suggest the potential importance of the hospital emergency room as a site for identification of family violence problems that may be precursors of homicide. Actions such as the following are needed: • Improved medical protocols need to be developed for identifying adult victims of domestic violence. Many of these persons are not identified as battering victims because they do not volunteer this information and are not questioned about possible battering by medical personnel. A model emergency room protocol for identifying adult victims of domestic violence has been developed for hospitals in the state of New York and can be adapted for use elsewhere (51). • Preventive interventions for victims of domestic violence should be introduced and tested in health care settings. A leader in this field has been the Harborview Medical Center in Seattle which has developed a comprehensive intervention model that addresses the needs of victims of spousal violence, child sexual abuse, rape, elder abuse, and stranger assault (52). Police disturbance calls In the Kansas city study of homicide referred to previously, data were collected on the number of times police had responded to a prior disturbance call at the addresses of the homicide victims or offenders. It was found 52 that during the two-year period preceding each homicide, the police had responded to at least one disturbance call at the addresses of approximately 90 percent of the homicide victims or offenders. They had responded to five or more such calls at 50 percent of the addresses (53). For many years, police were reluctant to make arrests on domestic disturbance calls and usually limited themselves to a warning of some kind. During the 1970s, many police agencies abandoned this policy in favor of efforts at family crisis mediation that used a mental health or social work model. More recently, a number of departments have shifted to a more vigorous arrest policy as a result of political pressure and a study which indicated that arrests were more effective in stopping domestic violence than mere separation of the contending parties or attempts to mediate (54). It would be premature to conclude from this experience that arrest is always the best way for police to handle domestic violence situations. Rather, what seems to be suggested is that efforts by police agencies, health agencies, and others to prevent spousal violence can be facilitated when arrest is viewed as an option that can be invoked as needed. Gang Violence High rates of minority youth homicide in the nation's largest cities are associated from time to time with violence that develops among rival youth gangs. A study of the five cities with the most serious gang problems in 1972-1974 found that these cities averaged a minimum of 175 gang-related killings a year (55). In 1984, more than 100 deaths were attributed to gang-related homicides in Los Angeles. The House of Umoja in Philadelphia provides an example of a successful, minority-directed program aimed at prevention and reduction of gang homicide. The program was initiated in 1969 by two inner city Black parents who decided to become personally involved after one of their sons joined a youth gang. As an initial step, members of the son's gang were invited to live with the family in the manner of an African extended family. In 1974 to 1975, after a period of continued increase in gang-related homicides, the House of Umoja spearheaded a successful campaign to reduce gang violence by obtaining peace pledges from some 80 youth gangs. An important outcome of this effort was the formation of a community agency, Crisis Intervention Network that has worked to prevent a resurgence of gang violence through communication with concerned parties and organizational efforts to combat the environmental and social conditions that foster gang violence (56). NATIVE AMERICANS American Indian tribes differ considerably in language, culture, and traditions. Consultation with tribal leaders is vital when planning or implementing intervention targeted toward American Indian groups. It is 53 very important that non-American Indian health and mental health workers involved in prevention programs understand tribal history and cultural traditions, be aware of tribal-specific health problems, and understand local community requirements. Effective injury prevention measures should include both preventive intervention programs specifically targeted at identifying and changing high risk environments and interventions intended to change behaviors such as alcohol abuse toward a more healthy lifestyle. A comprehensive program planned by IHS to combat injuries, death, and disability will include resources from the tribes, IHS, and the Bureau of Indian Affairs (BIA) to improve tribal health problems, alcohol abuse programs, and emergency medical services. The IHS plan includes: • Prevention activities in the form of health education, risk identification and working with other agencies responsible for law enforcement programs. • Rapid assistance to injury victims through improved IHS clinical services and emergency medical services including assisting communities to improve ambulance services. • Rehabilitation through clinical services, alcohol programs, and mental health and health education. • Providing medical personnel with training in emergency care skills. Suicide Intervention efforts to prevent suicide among American Indians and Alaska Natives need not focus solely on suicidal behaviors. Related public health, community involvement, and cultural approaches are also relevant to primary prevention programs. Any suicide prevention and intervention effort, however, requires tribal sanction. Some efforts suggested as effective in reducing the rate of suicide among American Indian youth include: • Use of elders to give support to adolescents jailed for minor offenses. • Suicide prevention centers on reservations where counselors and trained elders can identify problems and provide support to high risk adolescents. • High school programs for pregnant adolescents, designed to lessen the possibility of suicide attempts and improve self-esteem and confidence. • First offender program where families work out drug and alcohol abuse problems with counselors. 54 • Early intervention prevention programs designed as long range efforts to prevent suicides and child abuse by teaching parents to provide good infant care. • Identifying adolescents at high risk, e.g. those who have had family troubles, school problems, or drinking or drug problems, and training them to acquire skills that would give them purpose in their lives and enable them to become reliable members of the tribe (57). Studies of suicidal behavior among different tribal groups are needed to identify the significance of variables such as living on or off the reservation, alcohol, employment, and other behaviors that may identify high-risk subgroups. Unintentional Injuries Automobile accidents are the major cause of accidental deaths among American Indians. Other major sources of injury include falls, fires, and drownings. An analysis of injuries and their circumstances among the Hopi Indians suggested a series of injury prevention strategies (58). Falls are the major cause of hospitalized injuries. The many falls in and near the home suggest the need to identify and remedy hazards such as loose rugs, poor lighting, and lack of railings on stairways and paths. Installing railings on top of buildings can prevent spectators from falling from flat roofs during tribal ceremonies. Roofs of pueblos should be inspected by appropriate agencies and reinforced, if necessary, to support crowds of spectators safely. Use of safety devices such as seat belts and child restraints in motor vehicles, and helmets for motorcyclists should be strongly encouraged. Passengers should be discouraged from riding in the backs of pickup trucks because they are in danger of being thrown out by sudden starts and stops, when rounding curves, or hitting bumps. Although not health related, widening roads, paving shoulders, and providing guardrails are recommended to reduce injuries from single vehicle crashes and rollovers frequently reported on the Hopi reservation. Alcohol has been implicated in many injuries among American Indians, but further study of its role should be documented, based on blood alcohol determinations. Studies of injury patterns among tribes other than the Hopi would be useful in identifying specific tribal problems and potential solutions. 55 REFERENCES: CHAPTER IV 1. U.S. Department of Health, Education, and Welfare. Public Health Service. Healthy People: The Surgeon General's Report on Health Promotion and Disease Prevention. DHEW Publication No. (PHS) 79-55071. Washington, D.C, U.S. Government Printing Office, 1979. 2. U.S. Department of Health and Human Services. Public Health Service. Promoting Health/Preventing Disease, Objectives for the Nation. Washington, D.C, U.S. Government Printing Office, 1980. 3. Falaka, Fattah: Call and catalytic response: The House of Umoja. In: Violent Juvenile Offenders: An Anthology. Mathias, R.A., DeMuro, P., Allinson, R.S. (Eds.) San Francisco. National Council on Crime and Delinquency, 231-237, 1984. 4. Heifer, R.E.: A review of the literature on the prevention of child abuse and neglect. Journal of Preventive Psychiatry 2: 99-111, 1984. 5. Moore, M.H.: The bird in hand: A feasible strategy for gun control. Journal of Policy Analysis and Management 2: 185-195, 1983. 6. Shah, Saleem A., Roth, Loren H.: Some considerations pertaining to prevention, pp. 148-149. In: D. Glaser (Ed.) Handbook of Criminality. Chicago, Rand McNally, 1974. 7. Hawkins, Darnell E.: Longitudinal-situational approaches to understanding Black-on-Black homicide. Paper commissioned by the Task Force on Black and Minority Health, December 1984. 8. Petersilia, Joan: Criminal career research: A review of recent evidence. In: N. Morris and M. Tonry (Eds.) Crime and Justice, Volume 2. Chicago, University of Chicago Press, 1980. Fagan, J.A. et al.: Violent men or violent husbands? Background factors and situational correlates. In: D. Finkelhor et al. (Eds.) The Dark Side of Families. Beverly Hills, Sage Publications, 1983. 9. Loeber, R.: The stability of antisocial and delinquent child behavior: A review. Child Development 53: 1431-1446, 1982. Hartstone, E. , Hansen, K.V.: The violent juvenile offender: An empirical portrait. In: R.A. Mathias et al. (Eds.) Violent Juvenile Offenders: An Anthology. San Francisco, National Council on Crime and Delinquency, 1984. 10. Estimates are derived from a 1976 survey of a nationally representative sample of American families conducted by the Family Violence Research Program, University of New Hampshire. 56 11. Fagan, J., Wexler, S.: Crime at home and crime in the streets: The relation between family and stranger violence. Unpublished paper prepared at the URSA Institute, San Francisco, 1984. 12. Hampton, R.L.: Family violence and homicide in the Black community: Are they linked? Paper commissioned by the Task Force on Black and Minority Health, December 1984. 13. Alvy, K.T. et al.: Black Parenting: An empirical study with implications for parent trainers and therapists. Rockville, Maryland, National Institute of Mental Health, 1984. 14. Straus, M.A. et al.: Behind Closed Doors: Violence in the American Family. New York, Anchor Books, 1980. 15. NIMH grant ROl MH 40027: Physical violence in American families: A resurvey. 16. Reports of the Culturally-Adapted Parent Training Project can be obtained by writing to the National Institute of Mental Health, Center for Studies of Minority Group Mental Health, Rockville, Maryland, 20857. 17. Olweus, D.: Stability of aggressive reaction patterns in males: A review. Psychological Bulletin 86: 852-875, 1979. 18. Robins, L.N.: Sturdy childhood predictors of adult antisocial behavior: Replications from longitudinal studies. Psychological Medicine 8: 611-622, 1978. Kohlberg, L. et al.: The predictability of adult mental health from childhood behaviors. In: B.B. Wolman, (Ed.) Manual of Child Psycho- pathology. New York, McGraw Hill, 1972. 19. Dodge, K.A., Frame, C.L.: Social cognition biases and deficits in aggressive boys. Child Development 53: 620-635, 1982. 20. Hannerz, U.: Growing up male. In: D.Y. Wilkinson, R.L. Taylor (Eds.) The Black Male in America. Chicago, Nelson Hall, 1977. 21. Spivack, G. et al.: The Problem Solving Approach to Adjustment. San Francisco, Jossey-Bass, 1976. 22. Lochman, J.E. et al.: Treatment and generalization effects of cognitiv. behavior and goal settting interventions with aggressive boys. Journal of Consulting and Clinical Psychology. In Press. 23. NIMH research grant application ROl MH 39140: Prevention with Black preadolescents at social risk. 57 24. Robins, L.N.: Negro homicide victims—Who will they be? Transaction, pp. 15-19, June 1968. 25. Ford Foundation: Hispanics: Challenges and Opportunities. Working paper. June 1984. 26. Garibaldi, A.M. (Ed.): In-School Alternatives to Suspension: Conference Report. Washington, D.C, National Institute of Education, 1979. Garibaldi, A.M.: In-school suspension. In: D.J. Safer (Ed.) School Programs for Disruptive Adolescents. Baltimore, University Park Press, 1982. 27. Boesel, D. et al.: Violent Schools—Safe Schools: The Safe School Study Report to the Congress, Volume 1. Washington, D.C, U.S. Department of Health, Education, and Welfare, 1978. 28. Fox, J.S.: Violence, Victimization, and Discipline in Four Boston Public High Schools. Boston Safe Schools Commission, 1983. Seashore, K.S.: Boston Teachers' Views about Problems of Violence and Discipline in the Public Schools. Boston Safe Schools Commission, 1983. 29. Bauer, G.L.: Restoring Order to the Public Schools. Phi Delta Kappan, 488-491, March 1985. 30. Nicholson, G. et al: Safe Schools: You Can't Do It Alone. Phi Delta Kappan, 491-496, March 1985. 31. Baker, K.: Research Evidence of a School Discipline Problem. Phi Delta Kappan 482-487, March, 1985. 32. Prothrow-Stith, D.: Interdisciplinary interventions applicable to prevention of interpersonal vioence and homicide in Black youth. Paper commissioned by the Task Force for Black and Minority Health, January, 1985. 33. Nicholson, G. et al.: op.cit. # 30. 34. Roizen, J.: Alcohol and criminal behavior among Blacks: The case for research on special populations. In: J.J. Collins (Ed.), Drinking and Crime: Perspectives on the Relationships between Alcohol Consumption and Criminal Behavior. New York, The Guilford Press, 1981. 35. Goodman, R.A. , Loya, F. , Mercy, J.A., Smith, J.C, Allen, N.H. , Vargas, L., Kolts, R.: Alcohol use by homicide victims: I. Associations with demographic characteristics. Unpublished draft paper, 1984. 58 36. U.S. Department of Health and Human Services. Fourth Special Report to the U.S. Congress on Alcohol and Health. National Institute on Alcohol Abuse and Alcoholism, Rockville, Maryland, 1981. 37. Haberman, P.W., Baden, M.M.: Alcohol, other drugs and violent death. New York, Oxford University press, 1978. 38. Washington Post, January 10, 1985. 39. U.S. Department of Justice. Bureau of Justice Statistics: Dictionary of Criminal Justice Data Terminology. Washington, D.C, U.S. Government Printing Office, 1981. 40. Federal Bureau of Investigation: Uniform Crime Reports for the United States, 1983. 41. Barancik, J. I., Chatterjee, B.F., Green, Y.C, Michenzi, E.M. , Fife, D.: Northeastern Ohio trauma study: I. Magnitude of the problem. American Journal of Public Health 73: 746-751, 1983. 42. Police Foundation. Domestic Violence and the Police: Studies in Detroit and Kansas City. Washington, D.C, Police Foundation, 1977. 43. Barancik, J.I.: op. cit. # 41. 44. Whitman, S., Coonley-Hoganson, R., Desai, B.T.: Comparative head trauma experiences in two socioeconomically different Chicago-area communities: A population study. American Journal of Epidemiology 119: 570-580, 1984. 45. Cooper, K.D., Tabaddor, K., Hauser, W.A., Shulman, K., Feiner, C, Factor, P.R.: The epidemiology of head injury in the Bronx. Neuroepidemiology 2: 70-88, 1983. 46. Sumner, B.B., Interviewing persons hospitalized with interpersonal violence related injuries. Paper commissioned by the DHHS Task Force on Black and Minority Health, January, 1985. 47. Mann, C.R.: The Black female criminal offender in the United States. Paper commissioned by the Task Force on Black and Minority Health, January, 1985. 48. NIMH research grant ROl MH 37180 final report. Identification and intervention with battered women in hospital emergency departments, Philadelphia. Philadelphia Health Management Corporation, 1985. 49. NIMH research grant ROl MH 30868 final report. Medical contexts and sequelae of domestic violence, 1983. 59 50. Straus, M.A. et al.: Behind Closed Doors: Violence in the American Family. New York, Anchor Books, 1980. 51. State of New York. Department of Health. Hospital emergency department protocol: Identifying and treating adult victims of domestic violence. Health Facilities Series: H-48, August 19, 1984. 52. Clement, J., Klingbeil, K.: A comprehensive model for identification, assessment, and treatment of victims of violence in hospital settings. Paper commissioned by the Task Force on Black and Minority Health, January 1985. 53. Police Foundation. Domestic violence and the police: Studies in Detroit and Kansas City. Washington, D.C, Police Foundation, 1977. 54. Sherman, L.W., Berk, R.A.: The specific deterrent effects of arrest for domestic assault. American Sociological Review 49: 261-272, 1984. 55. Miller, W.: Violence by youth gangs and youth groups as a crime problem in major cities. Office of Juvenile Justice and Delinquency Protection, Washington, D.C, 1975. 56. Falaka, F.: op. cit. #3 57. Linkages: Suicide among American Indian Adolescents. National American Indian Court Judges Association. March 1984. 58. Simpson, S.C, Reid, R., Baker, S.P., Teret, S.: Injuries among the Hopi Indians, a population-based survey. Journal of the American Medical Association 249: No. 14, 1873-1876, April 8, 1983. 60 CHAPTER 5 SUBCOMMITTEE RECOMMENDATIONS The Subcommittee on Homicide, Suicide, and Unintentional Injuries submits the following set of recommendations to the Task Force for inclusion in its report to the Secretary. The recommendations advanced by the Subcommittee are based on data resulting from high quality research and thus represent the best available knowledge. The Subcommittee believes its recommendations are reasonable and within the purview of the component agencies of the Department of Health and Human Services. The Subcommittee urges DHHS to exercise its national leadership role in identifying and bringing the devastating health problem of violence to the attention of all sectors of society. The recommendations fall into the following categories: • Information and Education • Access and Utilization • Capacity Building in the Nonfederal Sector • Financing Issues • Health Professions Development • Leadership Work With Other Sectors • Research Issues • Data Issues INFORMATION AND EDUCATION 1. The Secretary should make a clear statement that homicide is a preventable public health problem and that high rates of homicide in the United States are unacceptable. 2. Use the convening powers of the Secretary and the Department to call together groups of Black, Hispanic, and Native American health leaders to address homicide and other violence problems in their communities and recommend remedial health strategies. 3. Develop health promotion messages and campaigns conveying the message that homicide is a preventable public health problem and that high rates of homicide and violence in Black, Hispanic, and Native American communities are high priority health problems. 61 4. Develop, disseminate, and evaluate educational materials and curricula to target groups on homicide and violence and the antecedents which are believed to lead to these problems. 5. Encourage and support early implementation of IHS initiatives working with Federal, State, and local agencies to disseminate health information relevant to intentional and unintentional injuries among Native Americans. 6. Continue the process by the IHS of assessing the community and tribal needs for a comprehensive Community Injury Control program, leading to the development of realistic goals and objectives for the reduction of deaths and injuries from accidents. A multi- disciplinary approach involving other parts of IHS is necessary. ACCESS AND UTILIZATION 1. DHHS should take a leadership role in encouraging health agencies and leaders to undertake comprehensive reviews at State and community levels of the current response of the health care system to homicide and related violence problems (e.g. child abuse, spouse abuse, rape, assault). 2. Drawing on these reviews, DHHS should assist the health sector in developing and implementing improved health care strategies at State and community levels to prevent homicide and assist victims of violence. Particular attention should be paid to the needs of Black, Hispanic, and Native American communities. (Examples include: shelters for battered women, better record keeping, followup of victims following emergency room visits, and victim referral to social service systems by police following domestic violence cases.) 3. Improve emergency medical services throughout the IHS service area. 4. Continue to upgrade both personnel and facilities responsible for emergency medical services throughout the IHS, including the expansion of a system of one-step emergency care walk-in facilities. 5. Identify particular problems of transportation and communications which may now be barriers to access to and use of IHS emergency care facilities, particularly among remote tribal communities. Work with tribal, state, and local governments to overcome these barriers. 62 CAPACITY-BUILDING IN THE NON-FEDERAL SECTOR 1. DHHS should assist Black, Hispanic, and Native American communities in developing action strategies for prevention of homicide and other violence. Strategies should involve health, human services, edu- cational, and criminal justice personnel in this effort. 2. DHHS should begin systematic development, testing, and dissemination of model programs suitable for use in community-based efforts to prevent homicide and improve delivery of services to victims of violence. Such models might incorporate promising approaches such as the following: • House of Umoja program in Philadelphia, which has reduced gang-related homicide in Philadelphia. • Services to victims of violence available through the Victims of Crime Act of 1984. • "Big Brother" program to act as role models for high-risk children from single parent families. 3. The Public Health Service should seek the assistance of the private sector, especially foundations, to establish community programs in efforts to prevent homicide and other violence. • The 20 Robert Wood Johnson Foundation programs of health care for high-risk adolescents are an important example of how the private sector can contribute to these efforts. 4. IHS should continue to work closely with American Indian tribal leadership and community leadership among Alaska Natives, offering technical assistance and special, targeted training programs to expand and strengthen their capacities to prevent certain high- incidence events, such as reckless and drunk driving and, where indicated, family violence. FINANCING ISSUES 1. DHHS shold undertake a study of reimbursement mechanisms related to providing health care for violence victims and to developing more effective health care strategies to prevent homicide and other violence. Particular attention should be given to reimburse- ment mechanisms that could be used to support improved health care efforts to prevent homicide and other violence in Black, Hispanic, and Native American communities. 63 2. As part of DHHS efforts to highlight violence and accidents as priority public health problems, provide for predictable steady growth for the support of research and demonstrations leading to the prevention of these problems. HEALTH PROFESSIONS DEVELOPMENT 1. DHHS should foster increased awareness and recognition among health professionals that homicide is a preventable public health problem and that high rates of homicide and other violence in Black, Hispanic, and Native American communities must be reduced. 2. DHHS should assist the health care professions to develop improved training in homicide prevention and in identification, treatment, and followup of victims of violence. 3. DHHS could encourage Schools of Public Health and schools for training health care providers to add injury (intentional and unintentional) control courses to their curricula. LEADERSHIP WORK WITH OTHER SECTORS 1. DHHS should work with the Department of Justice and other Federal agencies to develop a combined Federal strategy for the prevention of homicide and other violence and improved delivery of services to victims of violence. 2. DHHS should work with the Department of Justice and other appropriate Federal agencies and private organizations to develop a national strategy for the reduction of homicides and other causes of morbidity and mortality associated with firearms use. 3. Explore ways to reduce the incidence of violence on television with FTC and the television industry. 4. DHHS and IHS should work closely with the Department of Housing and Urban Development, the Bureau of Indian Affairs, and representative tribal housing authorities to upgrade the current fire and environ- mental safety standards for the construction and rehabilitation of subsidized reservation housing. 5. DHHS and IHS should work closely with the U.S. Department of Trans- portation, its appropriate state counterparts, and tribal leaders on highway safety; on ways to improve highway design, signage, and furniture; and on the development of effective educational, legislative, and regulation strategies to change driver behavior. 64 RESEARCH ISSUES Although the Subcommittee's role was to examine the excess deaths of minorities due to violence, it is a problem of such complex magnitude, the Subcommittee believes the knowledge and data base must be increased for all people. Major research needs include: 1. Studies to understand better the patterns and variations in risks for homicide and other violence within Black, Hispanic, and Native American communities. For example, a study could assess differential risks for homicide in Black or Hispanic communities that are associated with variables such as the following: income characteristics, employ- ment status, percent of population failing to graduate from high school, school truancy, severity of community alcohol and drug problems, presence of street gangs, community values, prevalence of high risk lifestyles, and ease with which handguns can be secured. 2. Improved studies of situational correlates of homicide and nonfatal assault. These studies need to be aimed at identifying high risk situations for which preventive interventions may be effective. 3. Studies of homicides in high risk populations previously treated in hospitals or attended to by other mental health or social service agencies and other medical care settings for injuries caused by nonfatal violence (e.g. child abuse, spouse abuse, rape, assault) and for other presenting symptoms. Studies such as these are needed to assist in development of risk profiles for homicide that can aid in identifying persons in need of preventive intervention. 4. Studies to improve understanding of correlations between family dysfunction, family violence, childhood conduct problems, delinquent behavior, and serious juvenile and adult violence. 5. Studies to assess the efficacy of collaborative efforts between the law enforcement and health sectors to reduce and prevent violence (e.g., police referrals of victims of violence to health services, mental health followup of domestic disturbance calls). 6. Development and systematic testing and evaluation of biological, behavioral, and social interventions aimed at prevention of homicide and other violence in high-risk communities. Interventions may focus on structural risks for violence (such as inadequate schools or unemployment), on family and childhood problems linked to risk for violence, on individuals at risk for violence, and on situations conducive to violence. 65 DATA ISSUES 1. Support CDC efforts to improve monitoring and surveillance of morbidity and mortality due to intentional and unintentional injury in the high-risk groups of Blacks, Hispanics, and Native Americans and in the general population. 2. NCHS should allow for identification of specific minority groups, especially Hispanic and Asian/Pacific Islanders, in collecting data on morbidity and mortality. 3. CDC, working with other parts of DHHS, should develop strategies for surveillance systems and information sharing among legal, social service, and health agencies and strategies for implementing and evaluating these information systems. 4. The National Institute of Mental Health should continue to fund periodic surveys of domestic violence in the United States using representative national samples and oversampling for Black and minority populations. 5. The National Center on Child Abuse and Neglect should continue to fund periodic surveys on the national incidence of child abuse and oversample as appropriate. 66 Family Violence and Homicide in the Black Community: Are They Linked? Robert L. Hampton, Ph.D. Department of Sociology Connecticut College New London, Connecticut and Family Development Program Children's Hospital Center Boston, Massachusetts FAMILY VIOLENCE AND HOMICIDES IN THE BLACK COMMUNITY: ARE THEY LINKED? During the past two decades the issue of family violence has been transformed from a private issue to a public problem. Several studies established violence in home, toward both women and children, as a widespread phenomenon and part of the way of life for many families (Straus, et al, 1980; Walker, 1984). This new knowledge of domestic violence not only challenged many of our previous assumptions about family life, but also raised new questions about our empirical knowledge of the causes and consequences of violence in general. Although there has been an increase in the literature on domestic violence, only a few studies have actively examined racial and cultural differences (Garbarino and Ebata, 1983; Lindholm and Willey, 1983; Hampton, in press). Few researchers have attempted to develop a more precise conceptual and empirical understanding of the nature, type and severity of violence within Black families. As a result, even though there is widespread recognition of the fact that violence within Black families is a serious problem, there has not been an attempt to study this issue within a larger context. This paper examines previous research in this area as part of an effort to provide a framework for understanding violence within the Black family and other forms of violence in the Black community. We will begin with a review of research on each of three subareas: child maltreatment, family violence, and homicide and will highlight the most notable common traits in offender attributes and correlates. We will also discuss the extent to which Blacks, as a group differ from other groups in our society. Finally, we will pose additional questions to explore further the possible links among these phenomena. Child maltreatment, family violence, and homicide are arguably among the most serious social problems in the Black community. The evident seriousness of these problems justifies the current high level of public and professional concern. As a result, the possible interrelationship among these phenomena is now under consideration. Some researchers believe that child maltreatment, family violence and homicide are specific and discrete dysfunctions. Others, however, consider them as points on a continuum. In the latter view some homicides could be classified as fatal child abuse or fatal family violence. At least it would seem that child maltreatment, family violence and homicide have many overlapping causes. At worst, there may be a causal path: maltreatment in the family of origin may lead to family violence in the family of procreation, 69 which in some cases can result in homicide. Between these extremes we might argue for reciprocal causality (e.g. interspousal violence leading to child maltreatment) or interaction effects (maltreatment in the family of orl^tation leading to family violence, in family of procreation when other factors are present). It is possible that families affected by homicide may have been victims of other maltreatment as well. Not only the coincidence should be established, but also the magnitude, meaning, direction and significance of any apparent association. Is there a direct or indirect link between child maltreatment, family violence, and homicide? If there is a direct or indirect relationship, which participants are most affected? Can adult family violence be predicted from knowledge about child maltreatment? Or is there some set of factors that act at the individual, group, community or societal level to cause maltreatment, violence and homicide? If so, what other factors influence these additional factors? These are only a few of the questions that need to be addressed. Race and Child Maltreatment A. General Issue Although the true prevalence of child abuse is unknown. nationally, reported cases of child abuse and neglect have, increased 146% since 1976. In that year, 416,033 child maltreatment cases were reported to the child protective service system (CPS). The number of reported cases has increased annually, with 1,007,658 reported cases in 1983, the last year for which data are available. This increase can be attributed to a number of factors including definitional variability and surveillance artifact. According to the 1980 census, Black children accounted for about 15% of all children in the United States. From 1976 to 1980 the proportion of child abuse and neglect reports involving Black children remained fairly constant at about 19%, but 1982 national reporting data show that Black children were the reported victims in 22% of all child maltreatment reports (American Humane Association, 1983). Examinations of the relationship between race and family violence have yielded mixed results. In the first large scale summary of national reports, Gil (1970) concluded that families reported for abuse were disproportionately drawn from the less educated, the poor, and ethnic minorities. Black children were overrepresented as victims of abuse. More recent compilations show a similar picture (Spearly & Lauderdale, 1983: Jason et al, 1982) . Lauderdale and his colleagues (1980) computed annual rates of 70 the occurrence of validated cases of abuse and neglect in Texas from 1975 to 1977 for Whites, Black and Mexican-Americans. Without controlling for social class, Blacks had the highest rates for all forms of maltreatment, followed by Mexican-Americans and Whites. In a follow-up study using the same data, Spearly and Lauderdale (1983) added to their earlier study by controlling for social class and community (county) characteristics. Once again their results indicated that Blacks had higher rates of child maltreatment. The study also found that the greater proportion of high-risk families alone in a given area is not responsible for increased rates of maltreatment, but when combined with the presence of a highly urban environment it may yield particularly high rates for Blacks relative to the majority population. The National Study of the Incidence and Severity of Child Abuse and Neglect (NIS) funded by NCCAN and completed in 1981, provides some valuable information on racial differences in child maltreatment (NIS Study Findings, 1981). Based on a stratified random sample of twenty-six counties in ten states, this study collected information on child maltreatment cases from both Child Protective Services (CPS) agencies and non-CPS agencies; the latter include hospitals, schools, law enforcement agencies and others. Of particular relevance here is the fact that non-CPS professionals were asked to identify whether or not they had reported each known case to a CPS agency. This question allowed researchers to ascertain characteristics of cases known to professionals but not reported and to compare reported cases with unreported cases. The NIS estimated that 10.5 children per thousand are maltreated each year in this country. This estimates reflects not only physical abuse, but includes other forms of maltreatment as well. In terms of ethnic group, incidence estimates are essentially the same for Black and White children (11.6 and 10.5 per thousand, respectively). The study also reported that contrary to the conventional wisdom there were no observed reporting bias due to race (Study Findings, 1981). Subsequent analyses of these data have challenged both of these findings. One of the shortcomings of the original NIS reports is the classification "White", combined data from "White not of Hispanic origin" and "Hispanic". The assumption was that most children identified to the study as Hispanic would have been classified as "White", according to Census Bureau race definition. in a secondary analysis of substantiated maltreatment cases taken drawn from the NIS data, Hampton (in press) has shown significant differences among Blacks, Whites, and Hispanics in demographic profiles and in type and severity of maltreatment. 71 As shown in Figure 1, Blacks had higher rates of physical neglect and physical abuse than Whites and Hispanics. This analysis based on over 4000 confirmed cases of maltreatment, weighted to be nationally representative found that Black maltreating families were generally poorer, more likely to be receiving public assistance and include single parent mothers with fewer years of formal education. The mothers were less likely to be employed fulltime and had larger families, (see tables 1 and 2). Controlling for social class reinforces rather than diminishes these observed group differences. As table 3 shows a number of differences with respect to type of maltreatment and caretaker problems associated with the maltreatment. Even though the NIS reported no apparent reporting bias due to race in overall study, there is evidence that professionals from at least one report source exhibited reporting bias. Hampton and Newberger (1985) have shown that hospitals tend to overreport Blacks and Hispanics and underreport Whites. For Black and Hispanic families, recognition of alleged child maltreatment almost assured reporting to CPS. B. Physical Abuse The studies cited above addressed general rates of child maltreatment. Even though discussions of physical abuse are included, it is important to take a separate look at this form of maltreatment. In a study of 4132 cases of child abuse reported to the Los Angeles County Sheriff's Department from 1975 to 1982, Lindholm and Willey (1983) report a number of significant differences attributed to ethnic status. Physical abuse was highest in Black families, with discipline most often given as the reason for abuse. Differences existed also in the types of physical injuries that the children suffered as a result of the abuse. Black children were more likely to be whipped or beaten and to receive lacerations or scars, whereas White children were more likely to receive bruises. Although Blacks constituted only 9.8% of the Los Angeles County population, they accounted for 23.8% of abuse victims. These data also showed that among Whites and Hispanics, males (especially fathers) were the most frequent perpetrators. Conversely, Black mothers were more frequently identified as alleged perpetrators. Given the number of female heads of household in the sample, the latter is not surprising. The results indicated that the abusers were more likely to be females in single-parent homes but males in stepparent (consensual or legal) families. 72 Blacks were also overrepresented in confirmed physical abuse cases in the NIS (Hampton, 1985). Compared to others, Black victims of assaultive child maltreatment were more likely to be in the 6 to 12 age group, live in urban areas, have mothers who had completed high school, and suffer more serious injuries. As in the earlier analysis, Black victims of assaultive violence were more likely to be in households receiving public assistance. Caseworkers report that caretaker stress was highly associated with physical abuse in among these families. More than half (52%) the Black victims of assaultive violence had injuries from a weapon (knife, gun, stick, cord, etc.). In comparison only 27.4% of White and 44.4% of Hispanic assaultive violence victims suffered injuries inflicted by weapons, (see table 4). Taken together both the Los Angeles and the NIS data indicate a higher rate of physical abuse among Blacks. The former study suggests that this abuse is related frequently to parental disciplinary practices. Differences in family structure (specifically, one-parent versus two-parent families) are also noted for there contribution to ethnic differences. Although no SES variable was included in this study, one can infer from other variables that a large proportion of these families were poor. The NIS Black victims were in fact poor by comparison to non-Black victims, although 37% of the mothers were employed fulltime. A notable difference exists between the findings from these two studies based on official reports and a study based on self-reports. Using data from a national probability sample of 2,143 families (147 Black) Cazenave and Straus (1979) report insignificant differences in attitudes toward slapping and spanking between Blacks and Whites as measured by a Violence Approval Index. Blacks were less likely to report having actually slapped or spanked a child during the year preceding the study. Black and White respondents, however, reported nearly the same rate of severe parental violence against children. Cazenave and Straus report that when income and husband's occupation are controlled for, Blacks are less likely to engage in child abuse. This finding is directly contrary to the general assumption that Blacks are more likely than Whites to condone and use physical punishment of children. The authors also indicate that the existence of social supports, particularly from kin, family, and neighbors, may have an independent effect on controlling family violence (1979). It is beyond the scope of this review to attempt to fully assess the differences between these data obtained through self-reports and data obtained from other reporting sources. 73 Each approach has inherent strengths and weaknesses. Among the major drawbacks of the National Study (Straus, Gelles, Steinmetz, 1980) include the fact that no single parent families were interviewed although we know that such families are heavily implicated in violence. Another major drawback is that the sample did not include families with children under age three. Family violence The Straus, Gelles and Steinmetz study (1980) is generally cited as the primary source of data on the prevalence and incidence of family violence. In terms of overall family violence, they found sibling-to-sibling violence more prevalent than spouse-to-spouse, parent-to-child, or child-to-parent relations. Parent-to-child violence was second in terms of both prevalence and severity. Severe sibling violence was much more frequent in families in which parents were violent toward their children and toward each other; occuring in 100% of such households compared to only 20% of households in which parents did not use violence toward their children or each other (Straus, et al 1980). In the same study, Cazenave and Straus (1979) found that Black respondents expressed more approval of couple slapping than White respondents. Black husbands were also more likely to have slapped their wives and engaged in severe violence against them within the study year. Although the rates were relatively low for both groups, the rates for Black husbands were three times greater than the rates for Whites. There were only modest Black-White differences in the prevalence of wife-to-husband violence. Black wives, however, were twice as likely as White to have engaged in severe violence against husbands. Interspousal violence was highly associated with other forms of family violence in this sample. Straus et al found that women who were victims of severe violence were 150% more likely to inflict severe violence on their children than women who were not. These researchers also found that children who were victims of violence from their parents were more likely to use violence against their parents. Among those who had been hit the most by their parents, 50% used violence against them, while less than one in 400 who were not hit by their parents were violent toward a parent. These data along with anecdotal information from battered women's shelters (Fagan and Wexler,1984) indicate that there may be multiple victimization in many households. Battered wives may frequently be involved in the physical abuse of children and some child and adolescent abuse may be the unintended result of violence between parents. 74 Summary The first section of this paper has covered some of the major empirical studies on child maltreatment and family violence. Many of these studies were concerned with issues of measurement or estimation of the incidence of intrafamily violence. It is difficult, however, to make a rigorous assessment of this literature because a number of methodological and substantive issues are associated with research on this "sensitive" topic. Our best guess is that whether researchers use "self-reports" or some type of "official records", the estimates obtained will still fall significantly below true levels of incidence. This research has addressed a number of important theoretical issues, but has generally discredited an intraindividual approach to domestic violence. The summary evaluation of the psychopathological approach is that the proportion of individuals who batter their family members and suffer from psychological disorders is no greater than the proportion of the general population with psychological disorders (Gelles, 1983). This research has also demonstrated a relationship between exposure to and experience with violence as a child and violent behavior as an adult (Gelles, 1980). While this relationship exists, it is more probabilistic than deterministic. Investigators have found a consistent relationship between stress and violence, but stress in and of itself does not cause violence. Environmental stress, like childhood exposure to violence may be necessary factors in a violent relations, they are not sole causes of violence. Simplistic, unicausal models are generally dismissed by most of those who work in this field. Official statistics on child abuse and wife abuse indicate that women, Blacks, minorities, and the poor are overrepresented victims of domestic violence. Research that is not limited to studying officially "labeled" cases of domestic violence also finds relationships between income and family violence (an inverse relationship), race and violence, and minority status and violence (Straus, Gelles, & Steinmetz, 1980). II.Homicides In this section we review research on the prevalence, distribution, and correlates of homicide beginning with studies on child homicide. Next we will discuss studies that have souqht to distinguish among types of homicide with respect to victim-offender relationship and in the final section we review some research which has attempted to explain the homicides and other violent behavior. A. Fatal Child Abuse 75 Homicides of children by their parents or other caretakers represents the most serious form of child maltreatment. Like other forms of maltreatment, child homicides are probably underrecorded by current health statistics (Jason, 1983; Jason, Gilliland and Tyler, 1983). Using FBI-UCR for 1976 thru 1979 Jason and colleagues examined rates for three different categories of child homicide. Neonaticide, infanticide and filicide refer respectively to the murder of children less than a week old, more than 1 week old but less than 1 year old, and more than a year old. During the four year period from which these data were drawn, Black children were victims in 43%, 45% and 38% of the neonaticides, infanticides and filicides. Parents were the offenders in more than 2/3rds of neonaticides and infanticides. Rates of homicide by white fathers or stepfathers were 10% greater than those by white mothers or stepmothers (an insignificant difference) and rates by Black fathers or stepfathers were 50% greater than those by Black mothers or stepmothers (P<.001). The proportion of parent-perpetrated homicides is inversely related to victims age. When the age of the victims reached three years, the majority of homicides were not committed by close relatives of the victim (Jason, Gilliland and Tyler, 1983) . The type of weapons used varied with sex of both victim and offender but not with their race. When use of weapons was stratified by the age of the victim, parents used a higher proportion of firearms than did other offenders. In 1979, the latest for which data were analyzed, Black children were victimized at a rate 3.7 times that of whites. This analysis considered only cases of law enforcement-recorded homicide; thus it excluded more subtle cases of homicide and concentrates on active, lethal violence (Jason, 1983) . These data frequently exclude child fatalities associated with neglect (Jackson, 1984). B. primary Homicides The Black-White differential in rates of homicide victimization is well documented (Munford, et al., 1976; Costantino, et al., 1977; Rice, 1980; Farley, 1980; Mercy, 1983; Jason, Flock and Tyler, 1983; Jason, Strauss and Tyler, 1983) . In his analysis of homicide trends, Farley (1980) found that among men at the ages of maximum risk of homicide, the rates for nonwhites are eleven times those of Whites while among women, homicide rates for nonwhites exceed those of Whites. Homicide is the leading cause of death of Black men and women aged 25 to 34. The racial profile of homicide victims changed during the 76 nine year period, 1970-78. In 1970, 46% of these victims were White, and 54% were Black or of other races, whereas in 1978, 55% of victims were White, and 45% nonwhite (Mercy, 1983). Despite the decline in the rate of homicide among Blacks, the rate was still 5.6 times higher than for Whites. Several investigators have pointed to the importance of analyzing homicides categorized on the basis of the victim -offender relationship or the precipitating circumstances (9mith and Parker, 1980; Jason, Strauss and Tyler, 1983; Jason, Flock and Tyler, 1983). Primary homicide, is the most frequent type, and generally involves family, friends or acquaintances (Smith and Parker, 1980). These tend to occur within the context of interpersonal relationships with intimates, and are often acts of passion (Mulvihill and Tumin, 1969). The second type, non-primary or secondary homicides, generally involves offenders and victims who have no prior relationship. They are generally committed in the course of another crime such as robbery or rape, but may include gangland slayings. Between 1976 and 1979, 63% of all homicide victims died from assaults not related to another crime. The 1979 rate for Black males was 7.3 times that for White males; the rate for Black females was 5.8 times that for White females (Jason, Flock and Tyler, 1983). In that year 6242 Whites and 5851 Blacks were homicide victims. Using FBI-UCR data for 1976-1979, Jason and colleagues report that Black female homicide victims were involved primarily with acquaintances (47%) and family (43%). Sixty-two per cent of Black male victims were involve in situations with an acquaintance and 20% with family. At all ages, the majority of homicides involved family or acquaintances; however, family involvement was least common, and involvement with strangers was most common, when the victim or offender was a teenager (1983). The higher proportion homicides involving friends or acquaintances among Blacks may be related to several issues. First it could reflect, the higher proportion of subfamilies and augmented families in the Black community. These types often include persons having no consanguinal or legal times sharing a household (Billingsley, 1968). Second, it could reflect the mounting rates of marital dissolution and high rates of single- parent families. In many of these cases, an ongoing relationship mav exist with another person which is substantivelysimilar to familial relationships, whether or not the inSivIIuals share a common dwelling. Third, these rates might reflect other aspects of the social context of the offender's lifestyle. The level and severity of family or acquaintance violence may be felt by family or friend, depending 77 upon the quantity and nature of time spent with each (Jason, Strauss, and Tyler 1983). In his analysis of homicide trends from 1960 to 1975 Farley (1980) reports that almost all of the rise in homicide mortality among nonwhites and a substantial fraction of the rise among Whites results from the increasing use of firearms. For nonwhite women, approximately 90% of the total rise in homicide came about because of the increasing frequency of firearm use. A less dramatic but similar trend is reported for nonwhite men. Farley concludes that if there had been no increases in firearm murders, the homicide rates for nonwhites in the mid-1970's would have been just about what they were in the 1960's. For Black offenders, handguns continued to be the weapon of choice in the late 1970's (Riedel, 1984; Jason, Flock and Tyler, 1983). Only slight differences are noted between Blacks and Whites in the distribution of weapons used in homicides. One important difference, however, is that approximately one third of Black female offenders used knives; this proportion was much higher than for any other offender group. These studies reinforce several important points: first, that Blacks have a high rate of homicide; second, that homicides are primarily intraracial; third, that the majority of Black victims were friends, acquaintances or relatives of the offender; fourth, that firearms were the weapons most frequently used in homicide cases. C. Homicides and Other Violent Behavior Three popular but competing models which seek to explain differentials or trends in homicide. The first model has been called a "deterrence model," argues that the certainty of punishment, especially capital punishment, minimizes homicide. Ehrlich (1975) argued that capital punishment has a negative effect upon the incentive to commit homicide. His investigation suggested that the elimination of capital punishment was related to a rise in homicide. One major weakness in this model has been acknowledged: it ignores socioeconomic and demographic variables (Loftin, 1977) . This model also overlooks the fact that most primary homicides are neither psychotic nor premeditated acts. Wolfgang and Ferracuti (1967) estimate that no more than 5% of all homicides are planned or intentional. Subsequent research has shown little support for the deterrence model (Loftin, 1977; Parker and Smith, 1979). That is, there is little useful evidence that raising the costs for violence by inflicting more punishment (or speedier or more severe or more certain punishment) results in less violence (Currie, 1985). 78 The second model relies on cultural variables to account for differences in homicide rates. The "subcultural of violence" thesis argues that certain segments of society have adopted distinctively violent subcultural values (Wolfgang and Ferracuti, 1967). These values purportedly, provide normative support for violent behavior, thereby increasing the likelihood that hostile impulses will lead to homicidal incidents. Black-White differences in homicide rates are explained in terms of differing value orientations. More recent research has partially discredited this model. These investigations have shown that a "structural poverty index," which combines several socioeconomic variables is a more powerful predictor of State homicide rates than either race or region (Loftin and Hill, 1974; Parker and Smith, 1979). "The subculture of violence" model fails to explain variations in primary homicide rates: these rates appear to be more highly associated with social structural factors, especially measures of poverty. Racial differences in socioeconomic status are striking and several studies suggest that poverty may be a more significant factor than race with regard to primary homicides (Loftin and Hill, 1974; Bowman, 1980; Smith and Parker, 1980; Riedel, 1984). On the other hand, social structural factors are less important in explaining variations in secondary homicide rates (Smith and Parker, 1980; Messner, 1983) . The third model, ecologic analysis, incorporates socioeconomic variables and looks at the correlation between these and other factors. Rose (1984) suggests that one should not totally disregard subcultural explanations, but look at subculture, SES, and urban variables to better explain Black-on-Black homicides. Within this model, one must first recognize that the Black community is quite diverse. This diversity manifests itself not only in economic life, occupational pursuits and network structures, but also in rates of antisocial behaviors. It is evident that risk of homicide victimization varies greatly within the Black community. The extent of variation can be related both to the nature of a given city's spatial configuration and to the sorting of population along SES and lifestyle orientations (Rose, 1984). Environmental attributes may be associated with inherent risk promoting activities that may operate independently of external forces. in his study of Black homicides, Rose found that within the Black community: a set of stable high risk environments could be identified. Nevertheless, the structure of 79 victimization was observed to differ within specific high risk environments...The presence of stable high risk environments seems to suggest the existence of a subculture of violence that is place specific. In those high risk environments where expressive violence represents the modal type, shapes of the southern regional culture of violence predominate; and the cast of participants generally involves person who have lost hope of ever escaping their marginal status. This research also found that attempts to explain risk at the neighborhood level, using only structural variables, yielded mixed results. An interaction existed between structural characteristics of a neighborhood and stress levels, such that high stress levels in a high-risk neighborhood were strongly associated with expressive violence (primary) violence. Structural variables were less likely to explain risk in those environments where stress was low to intermediate (Rose, 1984) . in addition, the dual labor-market system which perpetuates the Black population's economic marginality and the role of environment in value formation, places many Blacks in a position where they are likely to resort to violence either to preserve a valued position or relationship or to acquire some valued resource (Bowman, 1980; Rose, 1984). Ecologic models hold that homicides are the product of several variables operating at the macro and micro levels. This perspective emphasizes environmental factors for Blacks, including poverty, unemployment, substandard housing, stressful life events and conditions that may render individuals more or less vulnerable. When these factors are added to technological change—the rise in the supply of handguns—one can see at least a partial cause for the increase in Black homicide (Farley, 1980) . To date, few studies have attempted to further test and develop ecologic models for understanding resent homicides in general. During the most recent upsurge in homicide frequency, even fewer studies have devoted exclusive attention to Black Americans, the population at greatest risk of victimization (Rose, 1984). III. Attempts to Identify Possible Links Certainly no simple cause-effect relationship exists between family violence and homicide in the Black community; many intervening and confounding variables are probably present. Although family violence and criminal violence have been 80 investigated separately by sociologists and criminologists, few attempts have been made to integrate the emerging knowledge of violence in the home with other violence research (Fagan and Wexler, 1984). A. Childhood Exposure to Violence For both intrafamilial and extrafamilial violence, exposure to violence as a child seems to be an important precursor of adult violence (Gelles, 1980; Fagan and Wexler, 1984). Violence in one's family of orientation seems to increase the probability of violence not only in one's family of procreation, but also outside the family. In a comparison of the characteristics of spouse batterers with those of "generally" violent men, childhood exposure to violence was the strongest predictor of involvement in both intra- and extrafamilial violence, explaining 26% of the variance characteristics of spouse batterers compared to "generally" violent men, (Fagan, Stewart and Hansen, 1983) . These data support a "social learning" theory of violence. Fagan et al., suggest that even though there is strong support for the notion of the intergenerational transmission of violence, future research needs to examine the environmental properties that provide such reinforcement and the cognitive processes associated with violent responses in certain situations. Equally important is an understanding of the factors which enable abused children to avoid violence as adults (1983) . B. Homicides in Response to violence Many victims of child or spouse assault respond to the assaultive treatment by striking back at the perpetrator. Brown's (in press) comparative analysis of women who kill their abusers and battered women who do not indicates that the significant variables that differentiate these groups describe the man's violent behavior not the woman's. Here we see that family violence and homicide are directly linked. The frequency with which family violence leads to retaliative homicides, rather than leaving the abusive situation has not been well studied. It appears that frequently a link between family violence and homicide is forged when victims of maltreatment attempt to remedy their situation by retaliating. For example, according to Lenore Walker (1984) , approximately 25% of the women serving 30-50 year sentences at Renz Correctional Facility in Missouri are there for killing abusive Sates! She cites other research conducted with female Prisoners which indicate that they were being abused at the time they committed other offenses. This female prison population is disproportionately Black. 81 C. Homicide as Family Violence Primary homicides and family violence do seem to bear a family resemblance. Both appear to be in part products of high-stress, low resources and multiple-problem families. In a legal sense the distinction between family and friends may be more or less artificial. In reality, however, we are dealing conceptually with violence and homicides among intimates. Studies of courtship violence and violence among cohabitants support the broader perspective on violence among intimates. The specific details about battering yield a catalog of violent acts and injuries (Walker, 1984), including hitting, punching, kicking, striking with objects, choking, stabbing and shooting. Homicides can occur either as an anticipated or unanticipated consequence of any of these acts. D. The Social Ecology of Family violence and Homicide Fully developed causal models in this field are rare. While most researchers agree that violence and homicides are the product of complex multidimensional process, few have attempted to identify the linking variables at different levels of analysis. Garbarino's "ecological model" has been utilized to explain the complex nature of child maltreatment (1980; 1982) . The model considers the context in which violent interactions occur and might include questions of environmental (neighborhood or community) quality. It also assesses the cultural, political, economic and demographic factors that shape the quality of life for families. From this perspective and work in several other disciplines we have learned that the extent to which an individual is integrated into local structures of kinship, neighborhood, and community is related to their prospects of mental health, physical well-being, and violence (Currie, 1985; Comer, 1985). With its focus on the mutual adaptation of persons and environments, this model appears to have some utility for the study of other types of dysfunctional family interactions. An ecologic approach to studying links between family violence and homicide in the Black community seems to be a fertile area for additional empirical research. Conclu»lon§ Are homicides linked to other forms of family violence in the Black community? Our review suggest that they appear to be conceptually linked in that they may have a common etiology. Many factors identified by research to date appear to be associated with both processes. in cases when homicide occurs as a direct result of violence or in retaliation to violence, the links may be causal in both directions. 82 There are many questions the available evidence does not address, but it is clear that the prevention of primary homicides may be effected by our efforts to reduce the incidence of family violence. These efforts will require primary prevention techniques which might include educational programs for parents and children, stress reduction efforts through neighborhood support "teams", and family counseling or therapy. Secondary and tertiary prevention techniques will also require implementation. It is clear that band-aid approaches based on poor conceptual and empirical foundations will do little to reduce Black-on-Black homicides in the near future. We are challenged to further explore that complex set of factors associated with violence among Blacks and develop programmatic initiatives with solid financial supports to reduce the immeasurable costs of violence not only to the Black community, but to our entire nation. 83 SUBSTANTIATED CHILD MALTREATMENT CASES ? w o c 07 O Cl Physical I7T71 Black Ethnic Differences Sexual Type of maltreatment EST White Emotional nent V//A Hispanic Neglect TABLE 1 ETHNICITY BY CHILD AND MATERNAL CHARACTERISTICS (Substantiated Cases) Variable White Age of Child u^5 31.5% 6-12 50.4% 13-17 17.8 Sex of Child MaTe 50.5 Female 49.5 Mother's Education** 0-8 years 13.6 Some H.S. 47.2 H.S. Grad or more 39.6 Mother's Employment*** Fulltime 29.5 Part-time 64.7 Unemployment 5.8 Mother's Age 19 or less 4.6 20-24 16.5 25-34 20.7 35-44 26.0 45 and older 26.7 Don't know 5.5 Number of Children 1-----------------19.0 2 28.3 3 22.9 4 14.8 5 or more 15.1 Father in Household Yes 56.1 No 43-1 Unweighted N Weighted N (3183) (364, 697) Black 34.5% 49.3 14.4 48.9 51.0 27.1 51.5 21.4 18.5 70.5 11.0 7.4 19.1 18.4 29.3 19.3 6.6 14.8 23.7 19.5 14.6 27.4 27.0 73.0 (746) (74,503) Hispanic 45.5% 40.8 13.7 42.8 57.2 34.7 40.4 24.9 27.7 65.6 6.7 5.4 21.4 13.3 28.3 23.9 7.8 20.2 30.2 27.5 10.8 11.3 56.0 44.0 (224) (20,008) Total 32.8%* 49.7 17.0 49.7 50.2 16.7 47.7 35.7 28.0 65.1 6.9 5.0 17.2 20.0 26.7 25.1 6.0 18.6 27.5 22.3 14.8 16.8 51.5 48.5 (4153) (459,207) *Percents are based on weighting and may not total 100% due to rounding CaseTwith unidentified ethnic status are omited from all tables. ★♦Excludes 1330 women for whom data were missing. ***Excludes 220 women for whom data were missing. 85 Table 2 ETHNICITY BY FAMILY INCOME, PUBLIC ASSISTANCE STATUS MEDICAID ELIGIBILITY AND CITY SIZE VARIABLE WHITE BLACK HISPANIC TOTAL Family Income** Less than $7000 $7000-14,999 $15,000-24,999 $25,000 or more 50.1% 39.9 7.5 2.5 74.6% 18.7 4.5 2.1 47.0% 41.0 10.5 1.5 53.7%* 36.8 7.2 2.3 Receiving AFDC*** Yes No 34.8 65.2 62.7 37.3 30.0 70.0 38.8 61.2 Medicaid Eligible**** Yes No 35.1 64.9 59.5 40.5 11.4 88.6 37.4 62.6 City Size SMSA over 200,000 Other SMSA Non SMSA 32.6 23.6 43.8 47.9 17.5 34.6 83.8 13.0 3.2 38.2 21.9 39.9 Unweighted N Weighted N (3183) (364,697) (746) (74,503) (224) (20,008) (4153) (459,203) *Percents are based on weighting and may not add to 100% due to rounding. **Excludes 139 cases for whom data were missing. ***Excludes 166 cases for whom data were missing. ★★★★Excludes 215 cases for whom data were missing. 86 Table 3 ETHNICITY BY TYPE OF MALTREATMENT AND PROBLEMS CONTROLLING FOR INCOME VARIABLE TYPE OF MALTREATMENT Low Income Families* Physical abuse Sexual abuse Emotional injury** physical neglect Other Families Physical abuse Sexual abuse Emotional injury Physical neglect FIRST CARETAKER PROBLEM Low Income Families Alcohol abuse Child rearing Caretaker stress Other Families Alcohol abuse Child rearing Caretaker stress Unweighted N Weighted N WHITE BLACK HISPANIC TOTAL 18.3% 17.3% 16.3% 18. 0' 4.4 6.8 7.9 5.0 12.3 4.5 13.8 10.6 40.3 53.2 49.4 43.6 32.3 65.4 31.9 35.2 10.6 7.1 16.9 10.6 11.1 3.0 9.2 10.3 33.9 21.5 21.9 32.2 16.1 17.7 38.6 22.7 26.7 18.4 14.8 20.6 38.9 17.6 19.9 34.0 17.9 12.7 37.2 11.8 12.3 45.5 19.2 17.3 38.3 17.5 12.9 37.9 (3183) (364,697) (746) (74,503) (224) (20,008) (4153) (459,207) Allies with incomes less than $7000 were included in the low income income category. Fifty three percent of the families were in this category. . **Emotional injuries include both emotional abuse and emotional neglect cases. 87 Table 4 DEMOGRAPHIC DISTRIBUTION OF ASSAULTIVE VIOLENCE CASES BY ETHNICITY VARIABLE WHITE BLACK HISPANIC TOTAL Age of Child 0-5 36.3 21.6 41.6 33.8%* 6-12 45.8 59.9 53.4 48.6 13-17 18.0 18.5 5.0 17.5 Sex of Child Male 55.2 42.9 41.8 52.4 Female 44.8 56.9 58.2 47.6 Mother's Education 0-8 years 4.8 17.5 12.6 7.2 9-11 40.7 50.1 62.3 43.1 12 years or more 54.5 32.4 25.2 49.7 Mother's Employment Employed Fulltime 30.1 36.9 20.0 30.8 Employed Parttime 63.4 53.9 67.7 61.9 Unemployed 6.5 9.3 12.3 7.3 Number of Children in Household 1 23.0 14.3 17.9 21.2 2 32.4 17.5 24.0 29.2 3 20.4 21.1 19.7 20.5 4 14.3 15.6 30.6 15.2 5 or more 9.9 31.5 7.8 13.7 Number of Victims 1 54.2 43.7 74.3 53.1 2 22.3 10.9 6.1 19.5 3 12.8 13.9 1.4 12.5 4 7.3 11.3 18.2 8.5 5 3.4 20.2 0.0 6.3 88 Father in Household Yes 68.7 48.7 83.7 65.7 No 31.3 51.3 16.3 34.3 Family Income Less than $7000 36.4% 43.7% 31.2% 37.5%* $7000-14,999 48.5 37.8 39.4 46.2 $15,000-24,999 10.5 11.5 24.6 11.3 $25,000 or more 4.6 7.0 4.8 5.0 Role of Mother in Abuse Maltreator 50.5 51.9 22.3 49.6 Not Involved 44.3 38.1 72.5 44.3 Don't Know 5.2 10.1 5.1 6.1 Severity of Abuse Serious 9.2 11.7 9.3 9.7 Moderate 57.6 42.8 75.7 55.6 Probable 33.2 45.5 15.1 34.7 Caretaker Problems Associated with Abuse Alcohol/Drugs 1678 12.0 8.2 15.6 Physical Disability 5.0 .9 0.0 1.2 Child Rearing 16.0 23.1 15.1 17.1 Emotional 9.1 13.1 6.3 9.6 Stress 32.3 40.3 51.6 34.5 History of Abuse 11.0 3.3 12.5 9.8 Other 9.9 7.4 6.3 9.3 Source of Report to CPS Law Enforcement 14.5 7.9 17.4 13.4 Medical Sources 12.2 22.4 12.7 14.1 Schools 20.3 25.2 52.2 22.6 Other 53.0 44.5 17.6 49.9 With Implement Yes 27.4 52.0 44.4 32.6 No 72.6 48.0 55.6 67.4 Unweighted N (734) (205) (37) (976) Weighted N (92,008) (21,654) (5,008) (118,671) *Percents are based on weighting and may not add to 100% because of rounding. 89 REFERENCES American Humane Association 1985 Highlights of office child abuse and neglect reporting. Annual Report 1983, Denver: American Humane Billingsley, Andrew 1968 Black Families in White America New York: prentice Hall. Bowman, Phillip J. 1980 "Toward dual labor-market approach to Black-on- Black homicide." Public Health Reports 95, 6:555-556. Brown, A. in Battered Women Who Kill press New York: Free Press. Cazenave, Noel and Murray A. Straus 1979 "Race, class, network embeddedness and family violence." Journal of Comparative Family Studies 10:281-300. Comer, James 1985 "Black Violence and Public Policy: Changing Directions." in Lynn Curtis (Editor) American Violence and Public Policy. New Haven: Yale University Press. Costantino, Joseph, Lewis Kuller, joshual Perper and Raymond Cypess 1977 "An epidemiologic study of homicides in Allegheny County, Pennsylvania," American journal of Epidemiology 106, 4:314-324. Currie, Elliott 1985 "Crimes of violence and Public Policy: Changing Directions." in Lynn Curtis (Editor) American Violence and Public Policy. New Haven: Yale University. Ehrlich, I. 1975 "The deterrent effect of capital punishment: A question of life and death." American Economic Review LXV:397-417. Fagan, Jeffrey and Sandra Wexler 1984 "Crime at home and crime in the streets: the relation between family and stranger violence." Paper presented at the National Institute of Justice Workshop on Family Violence as a Crime Problem. October, Washington, D.C. 90 Fagan, Jeffrey A., Douglas K. Steward and Karen V. Hansen 1983 "violent men or violent husbands? Background factors and situational correlates." in David Finkelhor, Richard Gelles, Gerald Hotaling and Murray Straus (eds), The Dark Side of Families. Beverly Hills: Sage Publications. Farley, Reynolds 1980 "Homicide trends in the United States." Demography 17, 2:177-188. Garbarino, James 1982 Children and Families in the Social Environment. New York: Aldine. Garbarino, James and Deborah Sherman 1980 "High-risk neighorhoods and high risk families: the human ecology of child maltreatment." Child Development 51:188-198. Garbarino, James and Aaron Ebata 1983 "Ethnic and cultural differences in child maltreatment." journal of Marriage and the Family 39:721-736. Gelles, Richard J. 1980 "Violence in the family: A review of research in the seventies." journal of Marriage and the Family, 42 (November):873-885. Gil, David 1970 Violence Against Children: Physical Child Abuse in the United States. Cambridge: Harvard University Press. Hampton, Robert L. 1985 "Violence Against Black Children: The Current Knowledge Base and Future Research Needs." Paper presented at the NIMH Staff College Invitational Conference on Research on violence in the Black Family and Community. September 27, Rockville, Maryland. Hampton, Robert L. in press "Race, Class, and Child Maltreatment. Journal of Comparative Family Studies. Hampton, Robert L. and Eli H. Newberger 1985 "Child abuse incidence and reporting by hospital: The significance of severity, race and class." American Public Health, Vol. 75, No. 1: 56-60. 91 Jackson, Aeolian 1984 "Child neglect: an overview." in Perspectives on Child Maltreatment in the Mid '80's, Department of Health and Human Service Publication number (OHDS) 84-30338. Jason, janine, N. Ambereuh, J. Marks and C. Tyler, Jr. 1982 "Child abuse in Georgia: A method to evaluate risk factors and reporting bias." American Journal of Public Health,72,12:1353-1358. Jason, Janine, Jeanne C. Gilliland and Carl Tyler, Jr. 1983 "Homicide as a cause of pediatric mortality in the United States," Pediatrics 72, 2:191-197. Jason, Janine, Melinda Flock and Carl W. Tyler, Jr. 1983 "A comparison of primary and secondary homicides in the United States American journal of Epidemiology, 117, 3:309-319. Jason, janine 1983 "Child homicide spectrum." American journal of Diseases of Children 137 (June): 578-581. Lauderdale, Michael, Al valiunas and Rosalie Anderson 1980 "Race, ethnicity and child maltreatment: An empirical analysis." Child Abuse and Neglect, 4:163-169. Lindholm, Kathryn J. and Richard Willey 1983 "Child abuse and ethnicity: patterns of similarities and differences." Spanish Speaking Mental Health Research Center, UCLA, Occasional Paper No. 18. Loftin, C. K. 1977 "Alternative estimates of the impact of certainty and severity of punishment on levels of homicide in American States." Paper presented at the annual meeting of the American Sociological Association, Chicago. Loftin, Colin and Robert H. Hill 1974 "Regional subculture and homicide: An examination of the Gastil-Hackney thesis." American Sociological Review 29:714-724. Mercy, James A. 1983 "Homicide surveillance, 1970-1978, MMWR 32, 2ss:9-13. 92 Messner, Steven F. 1983 "Regional and racial effects on the urban homicide rate: The subculture of violence revisited," Amerian journal of Sociology 88,5:997-1007. Mulvihill, D. J. and M. M. Tumin 1969 Crimes of Violence, Report of the National Commission on the Cause and Prevention of Violence. Washington, D.C. Government Printing Office. Munford, Robert, Ross Kazer, Roer Feldman and Robert Stivers 1976 "Homicide trends in Atlanta." Criminology 14, 2:213-231. Parker, Robert N., and M. Dwayne Smith 1979 "Deterrence, poverty and type of homicide." American journal of Sociology 85, 3:614-624. Rice, Dorothy 1980 "Homicide from the perspective of NCHS statistics on Blacks." Public Health Reports 95, 6:550-552. Riedel, Marc 1984 "Black-on-Black homicides: Overview and recommendations." in Daniel Georges-Abeyie (ed) , The Criminal Justice System and Blacks. New York: Clark Boardman. Smith, M. Dwayne and Robert N. Parker 1980 "Type of homicide and variation in regional rates." Social Forces 59:146-157. Spearly, James L. and Michael Lauderdale 1983 "Community characteristics and ethnicity in the prediction of child maltreatment rates." Child Abuse and Neglect 7:91-105. Straus, Murray, Richard J. Gelles, and Suzanne Steinmetz 1980 Behind Closed Doors, New York: Doubleday. Study Findings 1981 National study of the incidence and severity of child abuse and neglect. Department of Health and Human Services. Publication number (OHDS) 81-030326. Walker, Lenore 1984 The Battered Woman Syndrome. New York: Spinger Wolfgang, Marvin E., and Franco Ferrauti 1967 The Subculture of violence: Toward an Integrated Theory of Criminology. London: Tavistock. 93 Longitudinal-Situational Approaches to Understanding Black-On Black Homicide Darnell F. Hawkins Associate Professor Department of Sociology University of North Carolina at Chapel Hill Longitudinal-Situational Approaches to Understanding Black-on-Black Homicide Along with increased interest in the prevention of homicide among young black males (Health United States : 1980; U.S.Department of Justice, 1981; Centers for Disease Control, 1983) has come an awareness that despite years of public concern and scholarly research there is still much that we do not know about homicide as a social phenomenon. Further, despite our knowledge of the racial desproportionality of homicide, we know little about the patterning and causes of black-on-black homicide. This lack of knowledge is not due to an absence of research; rather it is largely a function of the type of research that has been conducted, including a selective inattention to certain aspects of homicide in the United States. In this paper I propose that this lack of in-depth knowledge of homicide poses a barrier to currently considered prevention efforts. I further propose that the integration of what social scientists call longitudinal and situational approaches to the study of homicide will help improve our understanding of homicide among blacks and other groups. What is currently known about homicide? Generally, it has been reported that arrest and victimization rates are highest for young males, urban residents, the poor, Southerners, and blacks. Even these facts are less well documented than one might expect given the longstanding public interest in this phenomenon. (See Hawkins, 1983, 1984, 1985.) For example Wilkinson (1984) reports that while the homicide rate in large urban areas is greater than that in smaller cities, the rural homicide rate in the Northeastern United States exceeded that found in large cities. There has also been considerable disagreement over whether the high rate of homicide among Southerners can be explained by their socioeconomic status or is correlated with other factors. It has been consistently documented that people tend to kill members of their own race. Some studies have shown that *<>re than 95% of all homicides are intraracial. That fact along with the disproportionately high rates of black homicide mean that a large proportion of all homicides in the United States are black- on-black Data from the 1983 Uniform Crime Reports provide an illustration of recent trends. Among all homicides involving one offender and one victim where racial characteristics were known the following was reported: (1) 46% of all cases were white-on- white (2) 44* were black-on-black, (3) 5* were black-on-white (4 2* were white-on-black, and (5) the remaining 3* involved all other racial combinations, e.g., those involving other race offenders and other race victims, other race offenders and black or white victims, and other race victims killed by blacks and whites. Clearly, given the size of the black population in the United States (12%), black-on-black homicide represents the most disproportionately distributed category. An understanding of the social processes leading to this excessive rate is vital to our understanding of homicide in the United States and also to the 97 success of currently proposed prevention efforts. The information on the sociodemographic characteristics of homicide offenders and victims described above is useful. It serves the purpose of informing us of the characteristics of persons at whom intervention efforts should be directed but tells us little (except implicitly) about reasons for high rates of homicide among these groups, i.e., about the social conditions and processes that lead to murder.1 Why do some groups have higher rates than others? Are there discernible social factors that account for these rate differences? Can any of these factors be manipulated in order to prevent or reduce homicide? The conducting of research that addresses these kinds of questions has always been the goal of investigators in the social sciences. That goal remains largely unrealized, however. It is the purpose of this paper to critique previous research and to propose a research design that has the potential to permit researchers to begin to answer some of these questions. In the words of a previous task force on violent crime, MIt is imperative that we discover what works and what does not." 2 Earlier Studies of Homicide The behavior characte The firs sociodem of quant between such as forth. statisti attribut 1979). identify Over the certain the firs latter. study of ho al science d rized by the t of these i ographic met itative anal crime and po crime rates A second res cs and inste es of select The goal of and describ years these academic dis t approach a micide and uring the 1 use of two s what migh hod. At it ysis sought verty utili for entire earch metho ad involved ed samples the latter e the crimi approaches ciplines. nd pyscholo other crime in social and ast fifty years has been principal approaches or methods. t be called an ecological or s inception in Europe this method to examine the relationship zing spatially aggregated data, countries, regions, cities, and so d avoided the use of aggregated the study of the personal of criminal offenders (Greenberg, group of researchers was to nal personality or disposition. have come to be associated with Sociologists have tended to favor gists and psychiatrists the These studies have be of studies documented the among certain groups and i approach, though not as co possible psychological att studies have resulted in ( of a link between mental d been rejected by most rece Yet, the search continues aggressive and homicidal b 1977), and also possibly g (Ellis, 1982). These appr from failing to consider f en valuable. In fact the first group disproportionate occurrence of homicide n certain geographic areas. The latter mmon as the first, has provided data on ributes of homicide offenders. Such or have reinforced) public perceptions isorders and homicide. This link has nt researchers such as Lunde (1970). for personality correlates of ehavior (Yochelson and Samenow, 1976, enetic bases of violent behavior oaches to the study of homicide, apart ully the larger societal dimensions of 98 homicide (e.g., poverty, social conditions, circumstances) also do not explicitly consider racial differences. While personal disposition and sociodemographic analyses of homicide have been popular among social scientists, other approaches have the potential to provide much more information about the social processes and conditions associated with homicide. One such approach involves the study of the situational correlates of the homicide act. Another approach takes a longitudinal view of the homicide offender (sometimes victim) and his actions and behaviors prior to the act of committing murder. Monahan and Klassen (1982), who provide an excellent review of situational approaches to the study of violent behavior, note that such approaches have been highly touted in recent years. They attribute the advocacy of such approaches to the fact that individual level analyses (personality studies) have produced disappointing results for both the anticipation and modification of violence. And of course, ecological research offers little explicit guidance (except at the most macro level) for predicting or preventing violence. Despite such recent popularity, Clarke (1983) reports that situational studies remain somewhat rare and have not been incorporated into theories of the causes of crime. Below I review both the situational and longitudinal approaches to the study of homicide and assess their potential value for improving our understanding of black-on-black homicide. Studies of the Situational Correlates of Homicide Wolfgang (1958) in his pioneering study of homicide in Philadelphia was among the first to consider what social scientists refer to as the situational correlates of the homicidal act. After providing the standard sociodemographic data on the characteristics of offenders and victims, he went on to describe features of the situation (circumstances, environment, milieu) of each act. Among the various factors considered were: (1) method and weapon used to kill, (2) temporal patterns: time of day, week, month, and year, (3) location of the act, (4) alcohol use, (5) motive of the offender, (6) relationship between the victim and offender, and (7) victim precipitation.3 Later, other researchers, especially social psychologists, expanded this list to include more attention to the interpersonal interaction aspects of the environment/situation that result in murder or other violent behavior. These include the behaviors of the victim and offender during that period of time immediately preceding the homicide. (See Nohahan and Klassen, 1982, for a review of these studies.) Monahan and Klassen (1982:293) have noted that few explicit or consistent definitions of situations or environments are provided by researchers. In addition, since homicide research has been dominated by the sociodemographic and individual disposition research traditions, many of the situational studies of crime cited by them are concerned with crimes other than 99 murder, e.g., other violent or property crime. Nevertheless, research on homicide, beginning with Wolfgang, has identified several findings regarding the situational or quasi-situational correlates of homicide over the years. A partial listing of these findings include: (1) The handgun is the weapon of choice, especially in recent years. (2) A large percentage of all cases involve victims and offenders who are related or otherwise well acquainted with each other. (3) Many homicides occur within a place of residence of the victim and/or offender. Others occur in other rather predictable social settings. (4) Most homicides occur during weekends and between 8 p.m. and 2 a.m. (5) A large percentage of both victims and offenders have been using alcohol and/or other drugs. (6) The most frequent "motive" results from an altercation of seemingly trivial origin between victim and offender. A second most common situation is the domestic quarrel. (7) In many instances the victim may "precipitate" his/her victimization by engaging in seemingly nondefensive acts of aggression.4 The value of this cataloguing of the situational correlates of homicide depends on how such information is used. For example, can we proceed to identify and describe situations in such a way as to be able to predict homicide? Can we alter aspects of the environment or situation so as to result in successful prevention of homicide or other violent behavior? Can we identify differences in the situational correlates of black- on-black homicide that may explain its disproportionate incidence in comparison to that found among other groups? Although Monahan and Klassen (1982:292) note that a growing body of theory suggests that situational approaches may be useful for predicting and controlling violent behavior, social researchers have not gone beyond a mere listing of situational correlates such as that provided above. In addition, Clarke (1983) reports that most of the relationships between various situations and crime, including violent crime, have not been shown to be causal.5 For instance, there is considerable debate over such questions as whether increased availability of handguns leads to higher rates of homicide (Kleck, 1979; Wright et al., 1981); the extent to which alcohol causes violent behavior (see Gary, 1980 and Collins, 1984); and why and how some altercations and arguments lead to homicide while others do not. Of course, some situations reported in earlier research are purely descriptive and are not meant to necessarily imply causation, e.g., the fact that most homicides occur indoors, within a res idence. 100 Clarke (1983:230) further emphasizes that not all of the situational correlates identified in previous research can be manipulated in the interest of reducing crime and even where manipulation is possible there is little firm evidence that reductions in crime would result. He alludes to various ethical constraints which I discuss in more detail in another study (Hawkins, 1985). On the other hand, he suggests that these concerns do not explain why criminologists have not more fully investigated the causal aspects of crime situations. I also note the general absence of such studies, especially for the study of homicide among blacks (Hawkins, 1983, 1985). Before considering the contribution such studies might make to our understanding of black-on-black homicide, let us review studies within another significant tradition—longitudinal research on homicide and violence. Longitudinal Studies of Homicide and Violent Behavior Just as social researchers have failed to fully examine the situational aspects of homicide, they have also paid relatively little attention to the study of homicide over time. Farrington (1982) says that such studies may involve measures taken at two or more points in time based on the same sample or they may consist of samples of different people taken from the same population. Most past studies of homicide are cross-sectional studies of a single population at one point in time. Follow-up studies are seldom conducted and if they are, they are conducted with a different population, e.g., a 1958 study of Philadelphia followed by a 1965 study of Chicago. Long term studies of the same population and the same individuals within that population are the ideal type. This kind of cohort analysis characterizes the informative study of delinquency by Wolfgang, Figlio and Sellin (1972). Longitudinal data may be obtained from respondents (e.g., criminal offenders or observers), from criminal records, or from both. The benefit of such studies is obvious. They provide much more information about the criminal offender and his victim(s) and are much more valuable than a one- time study for predicting behavior and for devising intervention strategies. Longitudinal studies of aggression in children and adolescents show considerable stability in such behavior over time. If a child was rated high or low on aggressiveness at an early age (8-10), studies have found that such information was useful for predicting whether he would be rated high or low at a later age (Farrington, 1982:180-181). Wolfgang, Figlio and Sellin (1972) found that boys who committed delinquent acts at an early age tended to continue such acts as they became older; however, some recent studies have questioned the idea that deliquent youths inevitably become adult criminals. Farrington (1982:182-188) reviews a number of studies which suggest that adult criminals, especially violent criminals, have a history of previous criminal offenses. This has been a consistent finding. Wolfgang's (1958) study reported that 66% of the homicide 101 offenders in Philadelphia had been previously arrested for crimes against the person, usually aggravated assault. About half of the victims had been arrested for such crimes. Farrington (1982:181) notes that left unanswered by investigators who report continued aggressiveness or crime over the life course of a child or adult is the question of whether such stability is the result of the environment or the disposition of the individual. There has been a tendency on the part of researchers to emphasize individual disposition. In some instances there is more basis for this conclusion than in others, e.g., Wolfgang, Figlio and Sellin compared brothers in the same families in their study of delinquency. In other studies the stability of violence alone provides the basis for assuming that the internal disposition of the child or adult provides an explanation. Environmental factors include socioeconomic conditions, the family setting, short-term precipitating conditions and many others. They could also include the response of agents of social control such as the police and institutions designed to correct behavior. For example, Farrington (1982:185) cites several studies which found that persons convicted of crime at an early age may be more likely to commit offenses later on than are persons who are not arrested and convicted. One study of adults showed that the probability of a subsequent conviction for violent behavior increased after each conviction. The researchers suggested that the number of previous convictions for violence should be taken into account during sentencing. Does the criminal record label the individual as criminal and thus lead to more deviant behavior and/or perhaps a greater likelihood of arrest in the future? Do institutions of social control (juvenile homes, prisons) teach crime? Was continued violence due to the ineffectiveness of intervention or was it due to some internal disposition of the individual? Whatever the answers to these perennial questions, research has shown patterns of continued violent behavior by individuals over time. Previous longitudinal studies of violence and aggression have not dealt specifically with homicide. Most are studies of aggressive behavior of all sorts, including criminal and noncriminal aggression. Others are studies of violent crime, but including such diverse crimes as robbery, rape, assault and other crimes against the person along with homicide. In the latter type of study homicide is treated as merely one form of violence and no effort is made to ascertain the temporal or causal relationship between nonhomicidal violence and homicide. That is, they do not provide explanations or causes for continued violent behavior that eventually culminates in homicide. Many are studies of children or adolescents. Below, I suggest ways in which such studies might be improved and used to begin to investigate the social processes underlying black-on-black homicide. 102 >> OJ >>«H . 1 rH O X bo 0 OJ 1 1 U p X X P o OJ "O CO OJ CO > Ol Ht) u OJ 0 u x CO "d o C rH| > OJ C X u cu u 0 OJ O T3 p t-i CO X •H •H O CO C 'H u •H -H| >J OJ CO P p •H p CU CO T3 CJ OJ OJ u - rH 0 U > > -H OJ O •H O CO mi (0 -p co oji oi-hi-o u O o o p p a x co •H CO P -0 OJ Bt-O-P > -H I*h X OJ p 3 X u O P c CO 3 CO •H 3 OJ 0 -P O >>-H COCO CO XI 0 p CT -PITJl COI o O (0 c p . CO P 0 T3 X O P 0 > -P OJ ■P ^ c co 00 X) COI c Ol CO CO OJ -HIjCI-h CO o 0 CO CO OJ z OJ C CO c 0 •H -H ,C • u co c -h OJ 3 bO J C a co c COI CO OJ X) Bl OJI B rH a 0 X OJ cm B 3J! OH OJ U bOH OJ O OJ bo p C rH 0 CO Ol OJ bfi CO Ol OIXJI o 0 p u O -H B bO O C CO B ■p CU C > Ch OJ P -P C -H ■H CO 0 CJ B ui u bo G U|-P| X OJ u r-i 3 p 0 u co oj OJ CO OJ o • 0 x c co co T» X -h OJ a 3 • H o O c CO O CO O CO <\ X oj >^.—ii a>i o M 2 CO o C B M C 0 T3 rH U oj o co a ■ H P C P C -H U ■P OJ XI r-H OJ|rH| „ 0 0 •H p o rH| B ■H| 431 Ol CO . p u u rH W rH ^ O X oj c OJ CO X TJ OJ C OJ OJ T3 U O OJ a COI bO •H >» Ul *H| OJ bos CO a CU -P -H 3 co cu B OJ •H O -H CO 0 CO CO C XJ OJ cj 73 a CO c u 61 c P 00 COI-PI >l P . . X a 3 X CO 00 CO -H u B rH CO CO O OJ OJ O •H T3 TO C P OJ o •HI •H H SI-CI CO OJ rH p CO 0 O P CO 0) OJ CO P i 0 CO bO c 3 X X CO 3 C CO 0 CO OJ 0 -Ol x> U •H| bOXJI rH CO •H CO rH «H CO OJ •H 0 0 CO O O P C h 3 1- U T3 J C 31 (h OJ CO M-H| OJI OJ - c OJ c p OJ CO C rH X O CO 0 CO CO CO CO O o OJ -H -P -H ■Ml CO > OJ CU Bl -*->| U X o c CO >»X 0 - 3 CO -H p CO CO •H CO O o o c U X > O OJ ■HI bo 0 X COI u O -H 0 rH p "C CO -P > CO a cumh OJ CO 00 OJ CO 3 01 0 0 '-H.Cl OJI U 3 c CO 0 rH X TJ -H -H OJ -c co p o OJ CO CO CO P ■H OJ u OJ B tj a B i-3l CO •H U 31 -PI U|rH CU P CO OJ P CO OJ J C O O rH p -o co to CO ■P CO T3 rH O CO u •H -H c > a coi co CU -H -a rH OJ > T3 co co -h a 0 X rH ►> firlHH 3 X (H OJ OJ > X OJ T3I 0 CO 0, OJI COI OJI c CO (0 >> •H CO J* •H a s u bO -P -P CO ■H OJ CO CO 3 P CO U P O W "0 CI •H X CO U| PI COI 0 s bo 0 c CO OJ 0 w co o oji c c CO B c c a CO -H r-^ 1 U C -H COI P OJ COI 31-h ■p 0 • H 0 X ■H OJ X CI OJ •H OJ "O P c OJ O O O u CO rH bO a o o CO -Q rH >JTDI COI -P co «p rH B •H rH -p B U -h OIX p CO OJ CO CO -P -H T-t a rH CO •H CO C •H -H r~II OJ CO Ol Ol CO X O 0 0 p 0 OJ U J p CO OJ 3 3 CO -P -P CO > O C 0 m -p a COI 3 ■p C CI PI 3 p T3 X CO U OJ X CU O OJ P OJ U bo B "O P CO CO CO c 3-HH U CO 0 • CI o< c •H OJI Ol -PI P c o 3 OJ .p CO OJ -H CO CO bO CU u c CO 3 3 •H OJ U TJ OJ rH Xi 2 Ol OJ T» 31 OJI COI-h cox (H p a co c CuX > OJ bo CO CO CO ■P +i «H T3 CO CJ 3 U -C > -H -H O 3 O ■P 0 y u bO COI Ol COI 0 OJ c C TJ OJ 3 p •H 3 0 bo u OJ • -P l—t c -o - OJ bo CO C P OJ COI c ■Ml J C CI OJI •H -H X •H C X rH CO rH O >> 0 •H OJ -P CO C X •H CT3-H C OJ O u CD COI O p ■H| (0 C X , CO X U -P X CO Ul OJI-H 3 T3 CO CO r-i > r-H -HI QBI Ol CO o* OJ p a 0 T3 OJ U T3 CO u O r-i co 0 CO CO ~ CO P OJ rH O C CI o 0 CO OJI Ul Ol CO OJ -P CO CO c c p C ft c x CO rH CO CO > •H C OJ X) c rH C CO -H 0 ■H| c C XI OJI Ul 3 X) C OJ u •H U CO •H CO OJ P OJ • CO > CO X bfi CO "-JX 5 p p ■H P -h 3 -0 bo -p p OJ H •H C 3 OJ rH •H ■pi c CO >» COI Ol OJ c c CO u c B O p CO OJ CO B co B p c C O X -H U U CO +J %4 r-t a CI Ml o U CO OJI OIX CO 0 •H 3 0 OJ u •H rH CO C CO 0 •H -H -H •H 0 C OJ -P CO CO o x •H 0 r-\ u cu fc «p -h Ol CU X X CO CO ^ rH O CO 0 E* 0 If) 0 0, then the percentage is positive (Haberman, 1978). The transformed coefficients are delayed in Table 4. In the first column of Table 4, we can now see that Hispanics, for example, experienced 47% less MATE homicides than the inferred norm, white BiacTs experienced 62% more such homicides. The higher percen ages iLong Asians for CRIME and GUN homicides show they differ sub tantially form the other groups in their relative homicide-specific odds. Likewise, 121 the uniformly low percentages among Hispanics set them sharply apart from the other groups. Another Look at the Victim/Offender Relationship. The one set of truly unexpected logit outcomes is the comparatively positive profile of homicide risk for Hispanics. Throughout the 1970s, the homicide rate for Hispanic males was consistently twice the size of the city's yearly male rate, and the group as a whole can be categorized as a high-risk homicide population. If, then, Hispanics are not strongly associated with MATE, CRIME, and GUN related homicides (though they had the lowest odds of being a victim of a non-REPRESSABLE homicide), what homicides are they particularly associated with, if any? To explore this issue, we returned to the basic components of the victim/offender relationship. From the original victim/offender responses, we constructed six mutually exclusive categories of homicides. The categories imply an intimacy continuum of relationships: MATE, (nonmate) FAMILY, FRIEND, ACQUAINTANCE, GANG (rival), and STRANGER. Table 5 displays the ethnic/racial distribution of these new categories. The figures reveal an interesting split: STRANGER homicides are the most frequent homicides for Asians and Anglos, and ACQUAINTANCE homicides are the most likely homicides among Blacks and Hispanics. Next, we executed the equivalent of five additional logit regressions, using the same determinants as before (see PROC FUNCAT, SAS, 1982). The difference here is that the five dependent variables were based on the odds log (pi/p) where £ is the proportion of homicides that were categorized as MATE and p. are the proportions of the remaining five victim/offender categories. The R for the combined logit model was .52 and the transformed coefficients for the ethnic/racial groups are listed in Table 6. These results now suggest that Hispanics experienced higher-than- average odds for FAMILY and ACQUAINTANCE homicides, and they were the most likely victims of GANG-related deaths. The coefficients also dramatic- ally show that the most probable relationship between Asian homicide victims and their assailants was overwhelmingly that of strangers. DISCUSSION Our analyses have focused on the research question of the presence and subsequent degree of association of our selected types of homicide and their relationship to the ethnic/racial status of victims. The importance of ethnic/racial status as a demographic variable has been demonstrated in most major studies of homicide in the literature. Similarly, the demographic variables of male sex and youthful age are implicated in these same studies. From a comparative perspective, our data confirm that male sex and youthful age are the more powerful as predictors, although the importance of ethnic/racial status is also clearly established. 122 One way of gauging the comparative importance of sex, age, and ethnic/racial status is to consider the Los Angeles homicide dataset as a sample of a larger multi-ethnic population (of the larger Los Angeles County area or of the greater five-county Southern California region). If inferential statistical techniques are applied, then only four of the 16 ethnic /racial coefficients that are presented in Table 3 would achieve significance. None of the comparable coefficients in Table 6 would achieve an alpha level of less than 5 percent (though several were less than ten percent). This is not the case for the age and sex coefficients. These estimates are quite stable. But the age and sex categories were not affected by the same cell size constraints as the smaller ethnic/racial categories. However, we believe that the ethnic/racial differences reported here are real, albeit somewhat less precise, for two reasons. First, the data represent population figures for a well-defined geographic area; and, second, the results are in agreement with other findings generated from the same dataset (Loya, et al. , 1984). One of the most stark ethnic/race results to emerge is the finding that Asians run the highest relative risk of being killed in a crime- related circumstance with a handgun where the relationship between Asian victim and assailant is that of stranger. In addition, Asian homicides were considered by police criteria to be rep res sable and, therefore, among the most preventable types of homicide. It would appear that the conditional risk for homicides among Asians are heavily weighed on the non-primary end of the coninuum. This specific pattern of victimization suggests that Asians are more vulnerable to crime-related events that may escalate to a fatal flash point. If Asians could avoid those situations and circumstances where a crime is being committed and then offer a little resistance to the criminal who is likely to be armed, is very certain that their proportion of homicides would be substantially reduced. Of, perhaps, greater interest is the apparent absence of primary homicides among Los Angeles Asians relative to other ethnic/race groups residing in the same urban area. Cultural mores that value group co- operation and resolution of conflict through co-coperation and resolution of conflict through compromise especially within the family are strong within traditional Asian ethnic groups. In addition, traditional Asian cultures are known to have stronger sanctions against homicide as evidenced by lower county-wide rates of homicide. The Asian racial group may represent a risk-free or very low risk population for primary homicide. Additional research attention focused on non-primary homicide is also necessary to more exactly determine the circumstances of criminal investigation. However, prevention programs utilizing educational approaches to increased public safety may have particular utility in Asian-American communities in the city of Los Angeles. Likewise, an increased law enforcement presence may be quite effective in the prevention of these non-primary homicides. By comparison to the proportionately greater involvement of Asian victims in non-primary homicide, both Black and Anglo groups seem more vulnerable to these fatal assaults within their most proximate of 123 relationships. While the most intense conflicts may occur among Anglo and Black consensual intimates, it would appear that most of these women victims and male offenders are less affected by social and legal sanctions against the expression of personal violence that ends in homicide. These primary homicides of consensual Anglo and Black intimates were also regarded by police as most unrepressable. The deterrence power of a recognizable law enforcement presence cannot penetrate the drawn curtains and locked doors of households where intimates resort to lethal means in prevailing over one another. Therefore, consensual mate homicides may be among the most difficult to prevent in any ethnic/race group where it is prevalent. Nonetheless, intervention efforts must be formulated and tested. For example, we suggest that strategies be developed to identify high-risk couples the first time their conflicts become known to others. Public education programs should direct their case identification efforts toward family members, friends, and the traditional help providers—police, emergency room personnel, clergy, social welfare workers, etc. These gatekeepers may then strongly urge or require, through legislative initiatives, that identified couples accept assistance in the resolution of their interpersonal conflicts as is mandated for some child abusers. Unlike Asian and Anglo groups, Hispanic homicides cannot be positioned at either end of the primary/non-primary continuum. Hispanic murders are distributed within the primary category, although the pattern of relationships are noteworthy. Hispanics were proportionately lowest on consensual mate homicides but highest on non-mate family relationships (e.g., parent-child, sibling, extended family). Results suggest that consensual mates, particularly Hispanic women, may be relatively "protected" from this subgroup of primary homicide, while Hispanic males, related by blood, are at a greater risk. A similar pattern emerged within non- family homicides. Hispanic men were proportionately more often victims of the more distal relationship of acquaintance and gang, rather than f r iend sh ip re la t ion sh ip s. It should be noted that the magnitude of difference is considerably greater for Hispanics as compared to Black and Anglos. If Asians were not victimized by strangers in crime-related circumstances, Hispanic stranger homicides would likely be the largest proportion of ethnic/race groups. Thus, it appears that Hispanic males are the sex group at proportionately greater risk of becoming victims of less proximate and, perhaps, less intense interpersonal relationships. Hispanic males may be responding to cultural sanctions against the expression of deadly violence toward women, particularly if they are friends. However, Hispanic men within extended families and in acquaintanceship relation- ships may be regarded as rivals in confrontations where threats of self- esteem can arise from ananswered conflict. Cultural norms may not strongly inhibit Hispanic men from becoming aggressive in these circumstances. Another relevant factor may be cultural mores which emphasize the importance of utilizing extended family in both the roles of kinship and 124 friendship. As a result, interactions with friends, acquaintances, and other non-family relations may be less emphasized as compared to Anglos and Blacks and, in turn, result in less opportunity for conflict. The importance of variation in culturally sanctioned behavior as moderated by socioeconomic status, acculturation level, generational status, and length of time in this country must be recognized. Sub- sequent research must attempt to address these issues. The ethnic/race results indicate that Hispanics skew the distribu- tion of homicides in the primary classification, while Asians and Anglos skew the distribution in the direction of non-primary homicides. The pattern of homicide for Blacks stands apart from Hispanics, Asians, and Anglos. Interestingly, Blacks are proportionately more often involved in both primary and non-primary homicides. The comparative risk pattern is, thus, grimly double-edged for Blacks. In other words, Blacks distribute themselves bi-modality on these types of interpersonally proximate and distal homicides. The comparative profile that emerges is of Blacks vulnerable to both intimate violence by a consensual partner or by another family member and also by interpersonally distal violence resulting from involvement in crime-related activities. To further compound the problem for Blacks, they are often victimized as a result of handguns—a most lethal means of death. Moreover, they are also likely to be involved in homicides considered unpreventable by law enforcement authorities. The elevated pattern of relative risk for Blacks on these types of homicides may be one important reason the homicide rates for Blacks are at least twice as high as Hispanics and more than three times higher than the other ethnic/racial groups in the study. Prevention and intervention efforts for Blacks overlap those suggested for Hispanic and Asian groups. It is clear that these efforts will require a more generalized approach because of the higher degree of proportional involvement of Blacks in all types of homicide. This research has exemplified the utility of investigating ethnicity/racial demographic factors of homicide in a diverse community such as Los Angeles. Homicide is not equally distributed across ethnic/racial populations and differential patterns of comparative risk for a variety of homicide types did emerge. Formulation of prevention and intervention efforts must recognize that homicide is an ugly end-point of violence that is multi- dimensional and, therefore, impacts its victims differentially. RECOMMENDATIONS Research 1. Homicide research should focus attention specifically on the ethnic/racial demographic variables prominent in each community or region studied. 125 The importance of specifically including Hispanics and Asians in the study population—groups traditionally not included in investigations of homicides—in addition to Blacks is established by this investigation. 2. The primary-nonprimary continuum of homicide conceptualization should be further defined to identify subgroups or variables that are important in the formulation of prevention and intervention strategies. (a) it is necessary to develop more precisely defined crime- related circumstances to determine more exactly the antecedents and determinants of victim and offender interaction. It is important to ascertain if the victim was engaging in a criminal behavior and was then killed as a result of that illegal activity or if the victim was a victim of the criminal behavior. (b) it is also necessary to increase the level of specification of the relationship between victim and offender in the area of non-family homicide. For example, friend, acquaintance, and associate categories of relationships require more precision in their specification to better determine the degree of intimacy or interpersonal proximity involved. (c) other similar precise classificaitons of circumstances surrounding homicide is necessary to better determine the chain of causality. Interventions 1. Increasing the police presence and assuming a concomitant increase in the efficacy of deterrents is likely not to impact homicide rates. 2. Decreasing the availability of handguns and increased prevention of crime is likely to have some impact on homicide rates. However, its greatest success will be in a rather small segment of the population, e.g., Asians. A more sizeable impact may be seen in Black and Anglo populations. 3. Formulation of intervention in spousal and consensual mate relationships is likelty to have a large impact on Black homicide rates. 4. Interventions focusing on family members other than spouses and consensual mates is likely to be most important in Hispanic, Black, and Anglo populations. However, the greatest impact may be observed in the Hispanic group. 5. Interventions directed at friend and acquaintance relationships are likely to affect all groups and have the greatest influence in Hispanics. 6. Prevention and intervention strategies should be formulated for specific ethnic/racial groups. The differences revealed by this study suggest that the form of intervention made by necessity vary consider- ably from one group to another. 126 REFERENCES Curtis, L. A. 1974. Criminal Violence: National Patterns and Violence. Lexington, Massachusetts: Heath. Gibbs, J. P. 1968. "Crime, Punishment, and Deterrence". Social Science Quarterly 48 (March): 515-30. Haberman, S. J. 1978. Analysis of Qualitative Data: Introductory Topics. New York: Academic Press. Jason, J., Strauss, L. T., and Tyler, C. W. 1983. "A Comparison of Primary and Secondary Homicides in the United States". American Journal of Epidemiology 177(3); 309-319. Jason, J., Flack, M., and Tyler, C. W. 1983. "Epidemiologic Characteristics of Primary Homicides in the United States". American Journal of Epidemiology 117(4); 419-428. Loftin, C. K., and R. H. Hill. 1974. "Regional Subculture and Homicide: An Examination of the Gastil-Hackey Thesis". American Sociological Review 39 (October); 714-24. Loya, F., P. Garcia, J. D. Sullivan, L. A. Vargas, and N. Allen. "Components of Change in the Rate of Homicide in Los Angeles: 1970- 1979". Presented at Annual Meeting of the American Public Health Association. Anaheim, California, November, 1984. Parker, R. N. and M. D. Smith. 1979. "Deterrence, Poverty, and Type of Homicide." American Journal of Sociology (November); 614-24. SAS User's Guide: Statistics. 1982 SAS Institute Incorporated, Gary, North Carolina: SAS Institute Incorporated. Wolfgang, M. 1958. Patterns in Criminal Homicide: Philadelphia. University of Pennsylvania Press. 127 TABLE I CRIMINAL HOMICIDE AND POPULATION CHARACTERISTICS FOR ETHNIC/RACIAL GROUPS IN THE CITY OF LOS ANGELES: 1970-1980 GROUP MGLO BLACK ASIAN HISPANIC one HOMICIDES POPULATION PERCENTAGE NUMBER PERCENT 1970 1980 1299 27.2% 59.(1 43.3% 2265 17.4 YJJ 17.0 89 1.9 4.6 6.1 1096 22.9 18.3 27.5 33 2.3 .4 1.0 4782 IDO.UK 100.07o inO.OT NOTE: The number of hcmicides are fcr the pericd 1970-1979; the ethnic/ RACIAL STATUS CF 168 HCMICIDES WAS NOT DETERMINED, 128 TABLE 2 PROBABILITY DISTRIBUTIONS OF SELECTED TYPES OF CIMINAL HOMICIDES BY SEX AND ETHNIC/RACIAL STATUS FOR EXEWLARY AGE COHORTS IN THE CITY OF LOS ANGELES: 1970-1979 SEX AND TYPE CF HOMICIDE AGE AND. STATUS MALE FEMALE 35-44 MATE NON MATE MATE NON MATE Anglo .089 .911 .488 .512 Black .206 .793 .519 .481 Asian .000 1.000 .000 .000 Hispanic .065 .934 .476. .524 Other .000 1.000 ,500 ,500 s* CRIfE NON CRIME CRIf€ NON CRIfE Anglo .617 .383 .696 .304 Black .404 .596 .556 .W5 Asian .875 .125 1.000 .000 Hispanic .467 .583 .696 .304 Other .000 .000 .000 .000 35-44 GUN NON GUN GUN NON GUN Anglo .552 .448 .397 .603 Black .537 .463 .'58 .562 Asian .840 .160 ,143 .857 Hispanic .547 .453 .423 .577 Other .333 .667 .000 1.000 25-34 REPRESS NON REPRESS REPRESS NONREPRE Anglo .496 .504 ,221 .779 Black ,530 .470 .270 .730 Asian .727 .273 .200 .800 Hispanic .680 .320 ,357 ,643 Other .750 .250 1.000 .000 TABLE 3 LOGIT REGRESSION FOR SELECTED TYPES OF CRIMINAL HOMICIDES IN THE CITY OF LOS ANGELES: 1970-1979 DEPENDENT VARIABLES MODEL MATE CRIfE GUN REPRESSABLE CONSTANT -3.284* -.486 -1.540* -.914* AGE 15-24 1.983* -.895* 1.121* 1.236* AGE 25-34 2.672* -.754* 1.185* .799* AGE 35-44 3.076* -.644* 1.176* .556* AGE 45-54 2.801* -.267 .965* .764* AGE 55-64 1.919* .304 .794 .735* AGE 65+ 1.286* .768* .183* .822* SEX -1.754* .074 .531* .892* ANGLO .119 .296 -.225 -1.099* BLACK .480 .518 .051 -1.227* ASIAN - .472 1.052* .652 - .996* HISPANIC - .634 .200 -.268 - .457 L2 (92.46) (80.53) (71.50) (75.25) * THE PROBABILITY OF THE COEFFICIENT IS LESS THAN .05, 130 TABLE 4 ESTIMATED RELATIVE RISKS, IN PERCENTAGES, FOR SELECTED TYPES OF CRIMINAL HOMICIDES BY ETHNIC/RACIAL STATUS IN THE CITY OF LOS ANGELES: 1970-1979 TYPE OF HOMICIDE STATUS NATE CRIPE GUN REPRESSABLE ANGLO 12.6J 34.5Z -20.2% -66.SE BLACK 61.6 64.7 5.2 -70.7 ASIAN -37.9 186.4 92.0 -63.1 HISPANIC -47.0 18.2 -23.6 -36.7 131 TABLE 5 PERCENT DISTRIBUTION OF CRIMINAL HOMICIDES BY TIE VICTIMS' RELATIONSHIP TO THEIR OFFENDER AND TIE VICTIMS' ETHNIC/RACIAL STATUS IN THE CITY OF LOS ANGELES: 1970-1979 STATUS RELATIONSHIP ANGLO BLACK ASIAN HISPANIC OTHER MATE 14.0% 17.2% 7.2% 5.4% 13.8% NON MATE-FAMILY 7.6 8.9 1.2 6.7 3.4 FRIEND 16.3 19.6 9.7 10.3 13.9 ACQUAINTANCE 12.1 32.9 16.9 32.5 17.2 GANG 1.0 2.6 4.8 16.0 6.9 STRANGER 40.0 18.8 60.2 29.1 44.8 100.0% 100.0% 100.0% 103.0% 100.0% TABLE 6 ESTIMATED RELATIVE RISKS, IN PERCENTAGES, OF CRIMINAL HOMICIDES BY THE VICTIMS' RELATIONSHIP TO THEIR OFFENDER AND THE VICTIMS' ETHNIC/RACIAL STATUS IN THE CITY OF LOS ANGELES: 1970-1979 STATUS RELATIONSHIP ANGLO BLACK ASIAN HISPANIC NON MATE-FAMILY 386.5% 415.3% 23.ES 1009.2% FRIEND -6.4 -24.0 183.3 13.1 ACQUAINTANCE 35.3 31.3 77.9 269.0 GANG -85.2 -98.9 -11.7 103.3 STRANGER -38.9 -96.2 84.4 11.1 NOTE: The proportion cf homicides between mates served as the denominator IN DETERMINING EACH OF THE RELATIVE RISKS REPORTED ABOVE. 133 Changes in the Criminal Homicide Rate of American Indians for the City of Los Angeles: 1970-79. A Research Note Fred Loya University of California, Los Angeles The Neuropsychiatric Institute Philip Garcia University of California, Santa Barbara Chicano Studies Department John D. Sullivan Pitzer College Luis A. Vargas University of New Mexico Children's Psychiatric Hospital Albuquerque, New Mexico Nancy Allen University of California, Los Angeles The Neuropsychiatric Institute CHANGES IN THE CRIMINAL HOMICIDE RATE OF AMERICAN INDIANS FOR THE CITY OF LOS ANGELES: 1970-1979 A RESEARCH NOTE American Indians are part of the ethnic variety found within Los Angeles* 2.9 million residents. In relative terms, the number of American Indians is still quite small; however, in actual terms, their current population size actually ranks quite high among cities with large proportions of American Indian residents. Notably, their numerical presence in the City is on the rise. Between 1970 and 1980, their numbers rose from about 8,200 to 16,000. Thus, like other minority groups within Los Angeles, their growth rate is much higher than the total growth rate for the area. American Indians are popularly viewed as a significant minority population. As a consequence, social problems associated with other minority groups are normally attributed to this group even though data on them is conspiciously absent. The possible errors inherent in this practice are obvious. For example, one of the City's most disturbing trends during the past decade was the fact that the crude homicide rate almost doubled; but unlike other minority groups in the City, the incidence of American Indian homicide victims during the period appeared to be comparatively low and relatively stable. Naturally, a comparison of crude rates of homicides can mask or exaggerate between-group dif- ferences. More detailed information is obviously needed before some summary statements can be made about the level and pattern of American Indian homicide victims who lived in large cities. This paper directly addresses two basic empirical questions con- cerning American Indian homicide rates in Los Angeles: (1) the question of whether changes in the age structure of the American Indian population have any effect on their yearly homicide rate; and, (2) whether change patterns in their sex and age-specific homicide rates were similar to the City's majority population — Anglos. METHODS AND PROCEDURES Our research topic is a neglected area of empirical study. The major barrier has been the absence of racial-specific homicide data. Our data comes from the official records of the Los Angeles Police Department. The data span the 10-year period, 1970-1979. One of the main coding tasks was to identify the homicide victims according to a scheme of six mutually exclusive ethnic/racial categories. American Indians, Anglos, Asians, Blacks, Hispanics, and Others. Of the over 4,850 cases, 98% were successfully categorized into the six categories. Only 15 of the cases were identified as American Indian homicide victims. 137 The second task was to derive appropriate population estimates for the years 1970-1979 so that our homicide data could be expressed as rates. The basis for our estimates were the Hispanic and racial figures provided by the 1970 and 1980 decennial censuses (U.S. Commerce Department, 1984, 1973a, 1973b, 1973c). The Anglo base figures were represented by non- Hispanic white enumerations and the American Indian base figures were represented by non-Hispanic American Indian enumerations. We assumed a mathematical model of linear change and then interpolated the age distribution by sex for the six groups in each of the interim years. Age specific rates are frequenty based on five or ten-year age intervals and they are usually expressed as persons per year. In our case, ages were collapsed into three intervals: 0-14, 15-44, and 45 or more; the homicide rates were based on five-year periods. This was done to smooth out the year to year irregularities that were an artifact of the small number of American Indian cases. Specifically, the numerators were the number of homicides for the years 1970-1974 and 1975-1979 and the denominators were the population totals for 1972 and 1977, respectively. The derived rates are expressed as per 10,000 persons. The focus of our analysis is the change between the 1970-1974 and 1975-1979 homicide rates. This difference is partitioned into three components: (1) the portion of the change attributable to changes in the age structure; (2) the portion due to changes in the age-specific rates; and, (3) changes associated with the interaction of age changes and changes in age specific rates over time. We refer to the three components as the age, rate, and interaction effects. The standard formula for decomposing two rates into these three effects is (H2 - Hi) - (P2i - Pli)H2i + s Angeles are not as susceptible to fatal violence arising from interpersonally close relationships as the other major ethnic/racial groups. Perhaps, the smaller numbers of American Indians in this large urban setting nested with 2.9 million residents acts to insulate them from more intimate violence associated with the relationship of victim to offender of con- sensual mate, other family members, and friendship. More research is needed to ascertain if, for example, cultural prohibitions forbide the expression of assaultive violence toward one's intimates or if minority groups who remain relatively less acculturated than Hispanics, for example, are less exposed to frustrating social forces. Anger and resentment that accumulate may be displaced on the intimate family and friends. As is the case of Hispanics in Los Angeles, it is necessary to better understand how American Indians avoid fatal violence in their nuclear and extended families and in their friendships, but are subject to victimization by acqua intances. We are very cognizant of the few cases of American Indian homicides that serve as our data base. When transformed into population parameters, this small number accurately reflects risk probabilities for a specific group in a well-defined geographical area. In addition, since these numbers do represent all known American Indian homicides in a large metropolitan city and there are relatively clear patterns of victimization, the cases reported on in this research may be reflective of patterns in other similarly populated urban areas. We believe our exploratory analysis should always be judged in this light. RECOMMENDATION Research 1. It is important to develop and maintain a surveillance system of homicides by ethnic and racial groupings. 2. As has been recommended in another report, it is important to more precisely determine and record the multiple dimensions of victim and offender interaction. For example, increased precision in determinations of categories of circumstances, motives, relationship, and reason are necessary. 140 3. The specific relationship category of acquaintance should be the subject of more intense research. Our data suggest that the overwhelming majority of homicides involve victims who are killed by acquaintances. 4. Attention should be focused on comparing patterns of homicide victimization in other urban areas similar to Los Angeles. These data should then be contrasted to patterns of homicide in areas (reservations) more densely populated with American Indians. 5. With the great diversity of tribes of Native American Indians, it is also important to ascertain similarities and differences among subgroups of Indian people. Interventions At the local level, Indian centers and community leaders should be made aware of the overrepresentation of acquaintance relationship in American Indian homicides. Public education programs may be useful strategies to consider. 141 BIBLIOGRAPHY Loya, F., P. Garcia, J. D. Sullivan, L. A. Vargas, and N. Allen. Components of Change in Rate of Homicide in Los Angeles: 1970 and 1979. Paper presented at the 12th Annual Meeting of the American Public Health Association, Anaheim, California, November, 1984. Loya, F., P. Garcia, J. D. Sullivan, L. A. Vargas, and N. Allen. The Relative Risks of Types of Homicide Among Anglo, Hispanic, Black, and Asian Victims in Los Angeles: 1970-1979. Paper prepared for the Task Force on Black and Minority Health. National Institutes of Health. 1985. U.S. Department of Commerce. 1980. Census of the Population, Census Tracts. Los Angeles-Long Beach, 1983. U.S. Department of Commerce. 1970. Census of the Population, Subject Reports, Persons of Spanish Origin. 1973a. U.S. Department of Commerce. 1970. Census of the Population, Subject reports, American Indian. 1973b. U.S. Department of Commerce. 1970. Census of the Population, Census Tracts. Los Angeles-Long Beach, 1973c. 142 The Black Criminal Homicide Offender in the United States Coramae Richey Mann, Ph.D. School of Criminology The Florida State University Tallahassee, Florida THE BLACK FEMALE CRIMINAL HOMICIDE OFFENDER IN THE United States Female offenders of any racial/ethnic group were virtually ignored as meaningful objects of study until the early 1970s when researchers began to examine women who break the law, and their processing by the criminal justice system. With the exception of a small number of studies, few of which are empirically designed (see e.g. Deming, 1977; Lev/is, 1981; Mann, 1981; Riedel and Lockhart-Riedel, 1984), the American Black woman offender has been largely ignored as a viable research subject. Even more appalling is the lack of empirical attention devoted to the Black female homicide offender despite the startling incidence of her criminal homicide behavior. Studies of criminal homicide that explore gender and race reveal the following rank ordering in frequency of arrests: Black males, Black females, white males, and white females (Sutherland and Cressey, 1978: 30). It has also been reported that Black females have conviction rates for homicide that are fourteen times greater than those for white females (Sutherland and Cressey, 1978: 30). Thus, as Riedel and Lockhart-Riedel (1984: 2) so accurately remind us, "We are confronted with a paradox at the outset of our inquiry: of the four race and sexual combinations (Black males, Black females, white males, white females), Black females have the second highest rate and yet there is almost no research on the phenomenon." This paper is an attempt to partially close that research gap. The Uniform Crime Reports (UCR), collected from the various jurisdic- tions by the Federal Bureau of Investigation, and published annually by the U.S. Department of Justice, are the most frequently utilized sources of arrest data available in the United States today, but unfortunately the UCR do not cross-tabulate the information by gender and race. There are many problems associated with UCR data usage (see e.g. Hindelang, 1974; Skogan, 1974), but for the purposes of this topic, it should be kept in mind that minorities, especially Blacks, are frequently undercounted in the dicentennial census, which makes race-adjusted arrest rates suspect. Nonetheless, the parameters of the female homicide problem may be suggested from an examination of aggregated data on race and gender. The latest available UCR statistics reveal that in 1983, 49.6 percent of the total arrests for murder and nonnegligent manslaughter were Black persons (U.S. Department of Justice, 1984: 187), while the percent female was 13.3 for this offense category (Ibid, p. 186). When we consider the fact that, according to the 1980 U.S. Census, Black Americans only comprise about thirteen percent of the U.S. population, and that Black women are only about eleven percent of the female population in this country, we are able to obtain a rough image of the extent of Black female killing. However, in order to add another dimension to the picture of the incidence of Black female criminal homicide, in addition to the few empirical studies of the phenomenon to be discussed below, common sources of data are found in prison studies. These studies must also be viewed with caution, since incarceration is the final step down the ladder 145 of the criminal justice system and numerous researchers have found that trip fraught with "institutional" or other forms of racism at every step (e.g., Staples, 1975; Pope and McNeely, 1981; Georges-Abeyie, 1984). The Glick and Neto (1977: 153) national study of women in jails and prisons has been widely cited for its offense statistics which reveal a disproportionate number of Black women incarcerated for murder (18.6 percent of the Black prison offenses) compared to other racial/ethnic sub-groups (whites- 12.9 percent, Hispanics- 8.6, Native Americans- 13.4 percent). Similarly, 1979 California prison statistics indicate that Black females comprised the highest number of total homicide offenders (41.8 percent) followed closely by white females (38.8 percent) (Mann, 1984). In the New York State women's correctional system eighty percent of the women in prison for homicide in 1976 were Black women (Mann, 1981). Finally, a comprehensive study of women in prison in Michigan (Figueria- McDonough, 1981: 96) reports that 71 percent of the incarcerated female homicide offenders in 1978 were nonwhite. Obviously victim survey reports do not include homicide data but they nonetheless reveal a higher crime rate for Black women compared to other American women. Black females were perceived to be offenders in 64 percent of the female victimizations in a recent study reported by Young (1980: 29). Despite the fallibility of these sources of criminal data, it is clear that Black women's criminal involvement in general, and homicide involvement, specifically, are much higher than that of white, Hispanic, Asian or Native American women in the United States. On the other hand, high homicide rates among American blacks are not paralleled by such rates among African Blacks (Shin, et al., 1977) nor among African Black women. In many developing countries female crime rates are rising (Adler, 1979; Messner, 1983) and "newly liberated" African women are no exception in trying out crime; but the women in African countries are more likely to commit property crimes than crimes of violence and female homicide rates are negligible (Ebbe, 1984). The recent arrest in New York of what may be the nation's first female contract murderer, a twenty-one year old Black woman, would undoubtedly astonish her African sisters (Sufrin, 1984: 25). Homicide is predominantly an intraracial event. As whites kill other whites, Blacks kill other Blacks, and in the case of Black female homicide offenders, the usual victim is a Black male. The murder of Black men has reached epidemic proportions witnessed in the highest victim rates of criminal homicide in the country, or 58.5 per 100,000 (1978). Homicide is the leading cause of death among young Black males age 15-24 years (Mercy, et al., 1983), that is slowly decimating the Black male population. The second highest victim rate is that of Black females at 13.2 per 100,000 (Riedel, 1984: 53). Clearly there is an urgent need to obtain a more accurate picture of the Black female homicide offender not only because of the seriousness of the crime and the lack of information about this subgroup of offenders, but also to enable us to understand the underlying factors contributing to the offense and hopefully, prevent such violent 146 actions. The Present Study The purpose of this paper is two-fold. First, following the "itera- tive search" method used by Kleck (1981: 783) in his study of racial discrimination in sentencing, a "comprehensive assessment of the published scholarly empirical research" on female criminal homicide offenders was undertaken through an examination of sociological, criminological, and psychological Abstracts. The references listed in each of the identified books and studies were then examined until all informational avenues were exhausted. Through this process, and from data generously shared by other scholars, every recent empirical study of female criminal homicide offenders is believed to have been located. Each study was then examined for data on black criminal homicide offenders and where possible these data were then extracted. The second objective of the present study was to collect current empirical data on the Black female offender population from four urban cities with the largest homicide rates in the nation: Atlanta, Chicago, Houston, and New York City. A statistical examination of the Uniform Crime Reports for the years 1979 and 1983 was undertaken to isolate the target cities for homicide data collection. These years were selected for analysis because the UCR crime trends are reported in five-year periods thus enabling an examination of any changes in the arrests for murder over this time span. The national murder rate was 9.7 per 100,000 in 1979, and 8.3 per 100,000 in 1983. All states that had a murder rate of 10.0 per 100 000 in 1979 (N= 15) were ranked according to their rates and the Metropolitan Statistical Areas (MSA) with the largest rates within these states were then ranked to determine the urban areas "J^th« J^"* . . incidence of murder. An identical analysis was made of the 1983 UCR data and ten states were found to fit the criterion The objective was to obtain MSAs with the largest numbers of Black female crim!^1flhom^, offenders and simultaneously select those that could provide a regional picture of the phenomenon. Thus, the aforementioned cities were identified. Personal contacts made with key law enforcement administrators in the target cities were as follows: the Commissioner of Public Safety (Atlanta), he Supe ntendent of Police (Chicago), the Chief of Police Houston) and he S Police Commissioner (New York City). Each verbally indicated preliminary acceptance of the research plan and cooperation with the study. A draft survey instrument was forwarded to each administrator and several foow-up contacts were made. Unfortunately, the time parameters Sf the study (3 months) and various personnel constraints on the target city law enforcement agencies prohibited collection of those data. 147 The Analyses of Previous Research Definitions According to Wilbanks (1982), some researchers use the terms murder and homicide as if they were identical. Both terms have also been utilized in this paper in part because the UCR arrest category including the offenders under study uses the descriptive term murder and nonnegligent manslaughter. Some of the studies to be described referred to murder, others to homicide, and yet others to criminal homicide, the term preferred here. Following Wi lbanks' (1982: 153) definitions, Homicide, the more inclusive term, refers to the killing of one human being by another. This term covers both criminal and noncriminal (justifiable and excusable) homicides. Murder is a criminal homicide that involves both intent and premeditation. Some states distinguish between first-degree (generally requiring both intent and premeditation) and second-degree murder (requiring intent but not premeditation). Manslaughter is a type of criminal homicide, but is not considered murder. Other types of criminal homicide that are not considered murder are vehicular homicide and negligent homicide, (italics in original) It should be made clear at this point that the variety of inconsistent definitions in the female homicide offender studies examined introduce the very problem with which Wilbanks and other criminal homicide researchers are concerned. In order to circumvent this source of euphemistic error, attempts were initially made to redefine each study's data to produce a uniformly defined data set for examination. However it was observed that if such definitions are rigidly adhered to, an already limited number of studies would be further reduced. Since most of the studies involved small numbers, lacked control subjects, and varied widely in the types of variables scrutinized, hopes for standardization in reporting diminished. Therefore the female homicide studies are basically maintained in their original formats. Empirical Studies of Female Criminal Homicide Offenders A total of fifteen empirical studies that focused on female homicide offenders were identified. These research efforts span a 26-year reporting period (1958-1984) and include three studies that describe national rates and/or general trends in male and female homicide (Shin, et al., 1977; Wilbanks, 1982; 1983a), three homicide studies in which Black female offenders are not identified (Ward, et al., 1969; Rosenblatt and Greenland, 148 1974; Biggers, 1979), five studies wherein Black female homicide offenders are discussed in conjunction with white and/or other offenders yielding data which made it possible to isolate the Black sub-sample (Wolfgang, 1958; Cole, et al., 1968; Gibbs, et al., 1977; Totman, 1978; Wilbanks, 1983b); two studies in which Black female offenders, while not separable, nonetheless comprised the largest proportion of the sample thus enabling these data to be cautiously utilized, (Suval and Brisson, 1974; Weisheit, 1984) and two studies exclusively concerned with Black female criminal homicide offenders (McClain, 1981, 1982). The fifteen empirical studies of female criminal homicide offenders are depicted in Table 1 following the data base typology suggested by Wilbanks (1982). Wilbanks' (1982) notable critique of the literature on murdered women and women who murder describes the five most common data bases used in studies of female participation in homicide as the Uniform Crime Reports (UCR), statistics from the National Center for Health Statistics (NCHS), city study data, prison studies, and "anecdotal" studies. Since the emphasis in the present study is on empirical data, works utilizing the anecdotal, or case method, are not included. The remaining four methods seen in Table 1 are pertinent to this study and, with the exception of the UCR, which has previously been discussed, are briefly characterized following the descriptions offered by Wilbanks (1982: 155-160). Similar to the UCR, the National Center for Health Statistics provides cross-jurisdictional data. The NCHS information, gleaned from death certificates, is published in the form of national homicide victimization rates by sex and age and also includes details of the homicide event; but lamentably contains no information on the offender. Thus, depth of information is sacrificed by the researcher for breadth of information. City and prison studies, however, present the reverse situation by offering in-depth, detailed material on individual offenders while simul- taneously risking generalizability and comparability to other offender populations, since they are limited to their particular jurisdictional of ta po samples. The Findings The three studies utilizing UCR and/or NCHS data (Shin, et al., 1977; Wilbanks, 1982 and 1983a) are not included in this analysis which focuses on characteristics of the female homicide offender, specificaly the black ?emale criminal homicide offender (Table 2) vi^im characteristics Table 3), characteristics of the offense by offender (Table 4) offense charac- teristics (Table 5) and criminal justice data concerned with the offender and her homicide offense (Table 6). Salient points included in these three major studies have been previously introduced in this paper and will be interjected, where applicable, throughout. 149 TABLE 1 Empirical Studies of Female Criminal Homicide Offenders Researchers Year Reported Data Base1 195fl '" UCR NCHS City Study Prison Study Time Period No. Cases Ho. Black t Black 2 Wolfgang Philadelphia California 1940-195? 93 93 100.0 Cole, Fisher and Cole 1968 California 1965 111 4P 43.2 Ward, Jackson, and Ward 1969 Minnesota 1963-1964, 1960;1964-19G6 179 N.A. N.A. Suval and Brlsson 1974 North Carolina 1969-1971 87 70 B0.5 Rosenblatt and Creenland 1974 Canada (Also hospital records) 1970-1971 24 N.A. Glbbs, Silverman and Vega 1977 Florida 1977 43 26 60.5 Shin, Jedllcka and Lee 1977 X X 1940-1974 N.A. N.A. N.A. Totman 1978 California July-Dec., 1969 50 13 26.0 Biggers 1979 Florida 2 years 32 N.A. N.A. McClaln 1901 Detroit, 1975 119 119 loo.n McCla1nJ Wilbanks Wilbanks 1982 1982 198* St. Louis, Atlanta, Pittsburgh, Houston, Los Angeles Atlanta, Detroit 1975 1963-1979 1900 9 N.A. ?,4124 N.A. N.A. 100.0 N.A. N.A. Wilbanks Wclshelt 1983 1984 Dade County (Miami, Florida) Illinois 1900 1940-1966, 1901-Sprlng, ly03 47-' 4f.n 28 336 59.6 73.0 ^ased on Wllbank's typology (19P2), "Anecdotal" studies, the fifth category Is not included. Black sub-sample Isolated from the total sample. 3lh1s sub-sample 1s from 1901 study by McClaln. ^Although 2,412 reported, figure? In Vllbanks1 Tabic l-B total 2,012. 5AHhough 47 reported, figures In Wilbanks' Table 1 total 46. Offender Characteristics Even a cursory scan of Tables 2-6 indicates the level of attention the researchers devoted to each defined substantive area. It is clear that the researchers, both individually and collectively, report more information on the female homicide offender and the offense, than the other groupings of data. Race The number of subjects studied in the twelve research efforts under examination range from nine Black female homicide offenders who were interviewed (McClain, 1982), to the most recent study, reported by Weisheit (1984) on 460 incarcerated female homicide offenders admitted to the Illinois state prison for women from 1940-1966, and from 1981 to the Spring of 1983, 73 percent of whom were black. With the exception of two studies (Ward, et al., 1969; Totman, 1978), in every study where the proportion of Black female offenders is known, Blacks tend to predominate among the homicide offenders. The Ward, et al. study which included only 25 percent Black women1 was a prison study undertaken in Minnesota and California, two states with relatively small numbers of Blacks in the general population. California was also the location of the Totman prison study which lists the data by victim (mate or child) a fact that may explain the low Black percentages found (27.8 and 21.4 percent respectiye- lv) The other study that presented the data in a categorized fashion is the*Florida prison study by Gibbs, et al. (1977) who divided the findings into information on first, second, and third degree homicides. White women tended to be incarcerated for first degree murder (85.7 percent), while Blacks were primarily incarcerated for second degree (65 percent) and third degree murder (75 percent). Age Click and Neto2 (1977: 109) report that almost 65 percent of the incarcerated women in their national prison study were under 30 years of age whereas about 40 percent of the general female population are in that ale category. This finding is corroborated by the Michigan female prison studv9 (Iglehart, 1981: 38) which found that over half of all women committed to prison from 1968 to 1978 were below age 30 with a modal age oH o 24 years. Click and Neto further report that the most frequent age distribution for both Black and white incarcerated women was 22-25 vlars (25 7 and 28. 6 percent, respectively). Although the data on age shown n Table 2 indicates a variety of reporting methods- mean median, anHange- clearly the female homicide offender is not a young offender. Wilbanks (1983b), for example lists a range of 25-44 years in his 151 Table 2: Offender Characteristics la Race (%) Alcohol Substance Abuse (%) Narcotics Study Black White Other Mean Age (yrs ) I.Q. Education Yes No Yes No Wolfgang (1958)' H=93 100.0 - " 32.fi (median) - - " " "" ™ Cole, et al. (1968) N=lll 43.2 41.8 15.0 37.0 (W) 35.4 (B) 88.8 (W) 80.0 (B) 9.6 yrs • ~ " Ward, et al. N=179 25.0? 64.3 10.7 - Total 90=71X 90 nonwhite ---5oT~90 (1968) - 60.02 40.0 2.0 98.0 Suval and Brisson (1974) N=97 80.5 19.5 30.0 90.0 9.0 yrs 61.1 38.9 1.3 98.7 Gibbs, et al. -("1577) N=43 1 degree 2 degree 3 degree 60.5 14.3 65.0 75.0 39.5 85.7 35.0 25.0 _ 33.0 108.8 94.0 92.7 no high school diploma 42.9 85.0 68.8 85.7 70.0 75.0 14.3 85.7 30.0 75.0 25.0 75.0 14.3 25.0 25.0 Totman N=50 Mate Child 27.8 21.4 55.6 71.4 16.6 7.1 35.1 (W) 30.6 (B) B 85.6 80.0 W 109.2 89.4 Years B W 9.2 1076 9.0 10.4 - - - Biggers TT979T N=32 - - w — 35.0 90-100 11.0 - - - HcClaIn (1962) N=9 100.c - - 32.8 - - 12.5 87.5 0.0 100.0 -3 OS CJ n> to Studies by Rosenblatt and Greenland (1974), McClain (1981), Wilbanks (?83h), and Weisheit (1984) lacked sufficient data for Inclusion in Table, but are described 1n text where appropriate, as dre additional social characteristics not included in table because of small N's. 1063 and I960 samples. description of Dade County (Miami) female homicide offenders which tends to coincide with the aggregated age range found in the present study of 30-37 years. The age range of black female homicide of offenders in the studies by Wolfgang (1958), Cole, et al. (1968) and the two studies by McClain (1981, 1982), one of which is a sub-sample of a larger data base, is from 32.6 (median) to 35.4 (mean) years. On the other hand, the two studies that clearly identify white female homicide offenders (Totman, 1978; and Cole, et al., 1968) list mean ages of 35.1 and 37 years, respectively, suggesting that the white offender group for this offense may be older than black women who commit homicide. Obviously one can only speculate on such an implication, but an interesting research question is posed for future study of this variable. Intelligence Iglehart (1981: 42) found that about half of the women in prison in Michigan had I.Q. scores below normal (0-89), strongly suggesting "that the functioning capabilities of the incarcerated females are severely limited." Her colleague, Figueria-McDonough (1981: 86-87) states that it is among homicide and assault offenders that a higher relative incidence of women with scores below seventy was found," with only 37 percent of the homicide offenders having I.Q.s that were normal or above. In her exami- nation of sub-categories of offenses, Figueria- McDonough suggests that I.Q varies directly with complexity of illegal responses; for example, onlv 34 percent of the manslaughter cases (about one-half of the homicide commitments) had normal or above I.Q.s, but 48% of the first degree female homicide offenders tested at that level. Six of the studies under examination here included some measure of intelligence, although they vary in their reporting methods. In the four prison studies where Blacks and whites may be compared, whites scored higher on the intelligence tests administered. While the Pandora's Box concentng the reliability and validity of these tests will no be opened here it should be kept in mind that such tests have been challenged on racial and cu tural grounds. Cole, et al (1968) report higher I .Q.s for he white (88.8) than for the Black homicide offenders (80.0), with both «oum testing below the normal range. Other instances of less han n^rma n elligence are seen in the Suval and Brisson North Caro «na -tudy (1974) where 69.6 percent of the women, most of whom were Black 80 5 oercent) had I.Q.s below 90 and the Totman study in California (1978 whch eports both Black (80.0) and white (89.4) ^k!lle{^an?* ack ma e killers (85.6) at that level. Ward, et a . (1969) in heir ^ samole found that 42 percent of the women had less than 90 I.Q.s, especially the nonwhite group where 60 percent were below that intelligence level. There appears to be some support for the crime-complexity/I .Q. relationship espoused by Figueria-McDonough (1981) in the data reported in 153 two of the studies of female homicide offenders. First, Gibbs, et al. (1977) found that the first degree murderers (primarily white) had the highest I.Q.s (108.8) while the second degree (94.0) and third degree murderers (92.7) who were mostly Black had successively lower I.Q.s. A similar finding by Totman (1978) reveals an inverse relationship between the complexity of the homicide and I.Q., if one can assume that killing a mate is a more formidable task than killing a child. The mate killers, Black and white, had 85.6 and 109.2 I.Q.s, respectively; while the child killers were below normal intelligence with 80.0 (Black) and 89.4 (white) I.Q.s The only researcher reporting I.Q.s for the entire female homicide offender group in the normal range (90-100) was Biggers (1979) in Florida; although Ward, et al. (1969) found that more than half (58 percent) of the 1968 California sample attained this level, and 29 percent had I.Q.s over 110. In sum, three tentative conclusions may be suggested concerning the intelligence of the female homicide offenders in the studies under exami- nation: (1) the majority of imprisoned female homicide offenders have less than normal intelligence, (2) Black offenders incarcerated for homicide tend to have lower tested intelligence, as measured by prison- administered I.Q. tests, than white offenders committed for the same offenses, (3) the I.Q. of the homicide offender varies inversely with the complexity of the homicide as suggested by Figueria-McDonough (1981). Education As seen in Table 2, four of the five studies reporting information on years of education attained, show that female homicide offenders are unlikely to finish high school, since the average number of years of school completed is nine years or less. Again, prudent speculations about these data are possible relative to race. It would appear that the Black offenders have achieved less years of education since, according to data from Suval and Brisson (1974), they comprise about 80 percent of the sample and the education level for the sample was found to be less than nine years. Support for such a position can be gleaned from the Totman (1978) findings where Black mate and child murderers are reported to have 9.2 years and nine years of total education, respectively. White female homicide offenders, on the other hand, while still not completing high school, do have higher education attainment levels in both mate killings (10.6) and child killings (10.4). In addition, Gibbs, et al. (1977), report that the predominately Black second degree murderers, 85 percent of whom had no high school diploma, and third degree murderers with 68.8 percent in this educational status, lag far behind the dominant white first degree murder offenders where only 42.9 percent of their group has less than a high school diploma. The supposition that Black homicide offenders have acquired less years of schooling than their white counter- parts is not far-fetched, since Glick and Neto (1977: 129) found 154 significant educational differences between white and Black female offenders in the national study. Incarcerated female offenders, despite offense type, have generally been found to have educational achievement below the national average for the population (Iglehart, 1981: 40) and thus are frequently described as uneducated. The homicide offenders in the studied populations addressed in this paper are no exceptions to these earlier findings. The notion that differentiating life circumstances led to these women taking others' lives strongly suggests that the status of being disadvantaged, poor, and/or members of minority groups must certainly play an important part in reducing their life chances, opportunities, and achievements, among them educational accomplishments. The oftimes intolerable life of the lower- income individual, particularly those bearing the additional stigma of racial minority status can lead to a number of deviant avenues when stress is introduced. Homicide is one reactive behavioral symptom of such a social condition; substance use or abuse are others. Previous Alcohol and Narcotics Use A few of the studies of female homicide offenders explored the incidence of alcohol or narcotic use and report mixed results. Keeping in mind the small sample, McClain (1982) reports little evidence of alcohol use (12.5 percent) and no indications of narcotics use among the nine Black women studied. However, the larger sample in the Ward, et al. (1979) study also found that only two percent of the homicide offenders, most of whom were white women (64.3 percent), admitted to narcotics use. Suval and Brisson (1974) offer further corroboration of minimal drug usage in their finding that only 1.3 percent of the, in this case, mostly Black (80.5 percent), female homicide offenders, used narcotics. An opposite result is reported by Gibbs, et al. (1977) in that 75 percent of both second and third degree female murderers had narcotics histories,while considerably more (85.7 percent) of the first degree cases did. S nee These women^ere imprisoned in Florida, often called the "drug capital of the world," perhaps the incongruency of this finding, compared to the other studies, is not that unusual. Alcohol usage among the Florida sample (see Table 2) was also sub- stantial higher than in the other three studies reporting on this topic, ^with the exclusion of McClain's study (1982), drinking seems to be the rule, and not the exception for female homicide offenders. Other Offender Social Characteristics There are additional particularities of the female homicide offenders studied ha? are not included in Table 2 because they are reported by only a few of the researchers (marital status, broken homes, homosexual activity, 155 sexual promiscuity, and unemployment status), but they are mentioned here to add another dimension to the portrait of women who kill. Marital Status Both the national prison study by Glick and Neto (1977) and the Figueria- McDonough, et al. (1981) study of women imprisoned in Michigan reveal a majority of unmarried inmates. Only 19.9 percent of the incar- cerated women in the national study were married, compared to the reported 60 percent of married women in the United States at that time (Glick and Neto, 1977: 113). Whereas 22.4 percent of white women were likely to be single, almost one-third of the Black women (31.0 percent) were in that status (Ibid., p. 114). Blacks were also most likely to have always been single, in contrast to whites who were most likely to have had serial relationships (Ibid., p. 115). The marital status of the Michigan female inmates did not vary by race in 1978, but in prior years (1968-1972) a higher percentage of whites tended to be married (Figueria- McDonough, 1981: 52). An opposite picture is reported by three of the studies explored here, all of which found extremely high percentages of either evermarried or women who lived in cohabitation, and small percentages of single women. Suval and Brisson (1974) report the highest number of single women (27.1 percent) in their North Carolina study. One of the remaining two studies, both of which are in Florida, finds only 12.5 percent single women among the female homicide offenders (Biggers, 1979). Of more interest is the reported research of Gibbs, et al. (1977) who include race and degree of murder and find very few single women: first degree- 14 percent, second degree- 10 percent, third degree- 12.5 percent. Perhaps it is not that curious to find these three studies reporting evermarried female homicide offenders ranging from 72.9 to 90 percent in contrast to national and Michigan findings on incarcerated women. It is possible that since they took place in southern states, the results reflect the moral and cultural mores of the "southern tradition" which places women in a secondary status and frowns upon behavior contrary to the double-sex standard. Such a hypothetical notion assumes more credi- bility when one considers that North Carolina and Florida are in the southern "Bible Belt." Broken Homes Contrary to an opposite expectation, according to three of the studies that approached the subject, female homicide offenders do not tend to come from broken homes. Both Suval and Brisson (1974) and McClain (1982), studies with either predominantly Black or all Black samples, respectively found that 62.3 and 88.9 percent of the women did not come 156 from broken homes. Ward, et al. (1968), in their 1968 sample which was largely white (64.3 percent), report 68 percent of the homicide offenders were also products of intact homes. Homosexual Activity/Sexual Promiscuity Two studies investigated homosexual activity and not surprisingly female homicide offenders tend not to include such behavior among their personal characteristics. Only 11 percent of the Ward, et al. (1968) sample had been involved in homosexual comportment, while Gibbs, et al. (1977) find that such conduct is more typical of the first degree murderer (42.9 percent), than the second (25 percent) or third degree (37.5 percent) female homicide offender. Nor are Black women who kill likely to be sexually promiscuous, according to McClain's 1981 findings, since 91.7 percent of the Atlanta sample denied such behavior. But the 1968 sample of Ward, et al. (1968), 64.3 percent of whom were white, admitted in more than half of the cases (59 percent) to sexual promiscuity. Obviously two isolated instances tell little about this phenomenon, particularly concerning racial differences, but they do suggest that more research on this subject should be undertaken, Employment Status McClain (1982) reveals a majority (66.7 percent) of the Black female homicide offenders studied were unemployed. This figure coincides with the recent study reported by Weisheit (1984) which included 73 percent Black offenders and a mean unemployment rate of 63 percent. These figures are somewhat higher than the unemployed percentages reported by Glick and Neto (1977: 135) on Black (55 percent) and white women (50 percent), prior to their incarceration. Considering these unemployment statistics, albeit limited to a few studies, and the other dismal social characteristics of this group of violent women offenders, it is not surprising when Cole, et al. (1968 reoorrsTpeJcent of the white and 90 percent of the Black female homicide offenders worked in unskilled occupations prior to their incarceration. Suval and Brisson (1974) reveal that all of their sample, 80.5 percent of whom were B ack! occupied such a status in the labor force When one considers the national pattern of minority unemployment, underemployment and overrepresentation in lower status jobs in conjunc ion with the similar lower employment status of women in America it is patently clear ?nat mack women are doubly discriminated against- for being minority group members and for being females. 157 Victim Characteristics Most of the studies examined included information on the homicide victim but it was limited to the two particulars included in Table 3: the gender of the victim and the victim/offender relationship. Despite the paucity of data available, several interesting observations can be made about female homicide offenders in general, and Blacks within this category, specifically. There is clear verification for the claim that homicide is an inter- sexual^ intraracial, and intrafamilial event. Every researcher included here reported males, usually adult males, to be the primary target of the homicide. The proportions of men as the victims of women who kill ranged from 61 percent (Ward, et al., 1969) to the 97.9 percent indicated by Wilbanks in Dade County (Miami) Florida (1983b). A comment is in order about the McClain (1982) findings which reveal slightly more than ten percentage points difference between male (55.6 percent) and female homicide victims (44.4 percent). These data are based on only nine cases from Atlanta and Detroit and could reflect idiosyncracies of the sampling, particularly since the larger six-city sample from which these cases were selected (McClain, 1981) include 85.5 percent male victims. On the other hand, there are indications that Black females tend to kill friends more than white female homicide offenders and these friends could be females. In analyzing the victim/offender relationships in Table 3, for example, it is noted that among the second degree murderers (who are mostly Black) in the Gibbs et al. (1977) study, 40 percent of the victims were friends and acquaintances, as were 34.2 percent of the total Black subjects in McClain's larger study and 33.3 percent of her sub-sample. Many of the victims in the larger McClain sample could conceivably be women since they are clearly in the sub-sample, evidenced in the fact that male and child victims are accounted for as 55.6 and 11.1 percent of the victims, respec- tively. Indications of the intraraciality of homicides are seen in the victim/offender relationships. It appears that most of the victims of female homicide offenders are either husbands, lovers, family members, friends, or acquaintances. The personal closeness intimated by these descriptions, in American society today, are usually limited to members of the same racial/ethnic group, even when allowing for the slight possibility of an interracial marriage. If the victim/offender relationship data in Table 3 are collapsed for explication, it is readily seen that the persons most likely to be killed by women are members of their families (husbands, lover, child, other relative). The proportions of this category range from 40 percent (family) reported by Gibbs, et al. (1977) in the second degree murder group, to the 98 percent (husband/lover, child) revealed by Totman (1978). Averaging across the studies of all "family" combinations, as defined above, indicate that 68.5 percent of the victims of female homicide offenders are found in this subgroup. 158 TABLE 3: Victim Characteristics (1n percentnnes) VICTIM CHARACTERISTICS Wei she it (1984) VICTIM/OFFENDER RELATIONSHIP Friend Other Adult Male Adult Female Husband/ Lover Wolfgang (1958) 86.5 13.5 20.7 Cole, et al. (1968) White Black - - - Ward, et al. (1969) 61.01 16.0 35.01 Suval and Brisson (1974) 84.0 11.0 - Rosenblatt and Greenland (1974) - - - Gibbs et al. (1977) 1st degree 2nd degree 3rd degree - Totman (1978) - - 72.0 Biqgers (1979) - - 78.0 McClaln (1981) 85. S? 14.5 49.7 McClaln (19P2) 55.6? 44. <' 55.6 Wilbanks (1PP3) 0 97.9' 2.1 61.7 45.0 Child Family Acquaintance Stranger Unknown 20.0 12.0 19.0 5.0 28.0 2.5 11.1 8.5 9.0 46.7 47.0 56.0 87.5 5.1 21.8 18.0 8.3 34.2 33.3 14.9 29.0 11963 and 1968 samples. ?Sex of victims only, since child victims not Identified. 3.3 33.0 31.0 8.0 71.4 14.3 14.3 40.0 40.0 20.0 56.3 18.8 25.0 12.5 8.5 8.5 11.0 7.7 20.0 4.2 cr i—■ CO 9.4 6.4 3.0 TVk> of the twelve studies reporting victim/offender relationship consist of all Black female homicide offenders. These investigations reported by Wolfgang (1958) and McClain (1981) are separated by over two decades yet reveal some rather intriguing similarities. Both researchers collected city data with similar sample sizes: Wolfgang from Philadelphia (N = 93); McClain from Detroit, St. Louis, Atlanta, Pittsburgh, Houston and Los Angeles (N = 119). Wolfgang and McClain also report almost identical victim gender characteristics in that males were the most likely victims (86.5 and 85.5 percent), respectively. The two Black female homicide studies diverge, however, when victim/offender relationship is examined. Whereas in 1958 Wolfgang reports 67.4 percent of the homicides are "family affairs," (husband/lover, family), only 57.3 percent (hus- band/lover, child, family) of the victim offender relationships in the McClain study are so identified in 1981. A closer inspection of the two sets of statistics indicate that only 20.7 percent of the victims in the Wolfgang study were husbands and lovers of the offender and no children were victims, yet other family members were (46.7 percent). This config- uration seems to have changed by 1981 when McClain reports 49.7 percent of the homicide victims are husbands/lovers, 2.5 percent are children, and only 5.1 percent are family. Such drastic changes suggest that Black women are possibly becoming more violent towards those closest to them, or those whom they most love. It is also conceivable that Black women are simply becoming more violent. Some support for the latter position is suggested by an examination of the Wolfgang and McClain data concerning friend/acquaintance and stranger homicides, both of which show impressive percentages. Wolfgang's sample of Black female homicide offenders killed their friends or acquaintances in much lower proportions (21.8 percent) than the women in McClain's sample (34.2 percent). Further, stranger victims were reported by Wolfgang at only 3.3 percent compared to 8.5 percent by McClain. Another potential indicator of increasing violence among women may be seen in a perusal of the "stranger" column in the victim/offender portion of Table 3. Strangers were only 3.3 percent of the female homicide offender victims reported by Wolfgang in 1958; but by 1968, Cole, et al. are finding ten times that incidence for stranger/victim homicide among white females (33 percent) and 31 percent for the Black female group, in California. The predominantly Black second and third degree female murderers identified in the Florida study by Gibbs, et al. (1977) also have exaggerated percentages of stranger victimization (20 and 25 percent, respectively) and since Wolfgang, not one of the researchers, has uncovered a stranger/victim proportion less than two and one-half times the Wolfgang percentage for this victim category. Jason, et al. (1983: 310) differentiate between primary homicides or those that do not occur during the commission of another crime, and secondary homicides that occur during the perpetration of another crime. Most secondary homicides are intraracial (75 percent), but less so than primary homicides (95 percent). From 1976 through 1979, 19 percent of all homicide offenders were involved in secondary homicides. They were most likely to be Black (40 percent), but nonetheless, 31 percent were white. 160 Also, only five percent of secondary homicide offenders were female. Jason, et al. (1983; 316) constructed victim-specific percent distributions which reveal that the most frequent victim-offender/stranger relationship occurs in the secondary homicide classification. Using these data as a guide, extrapolations suggest that women may be demonstrating increasing violent behavior witnessed in higher proportions of stranger homicides possibly committed during the commission of another crime. Moreover, since a number of the female homicide studies indicate large percentages of Black offenders, with concomitant elevated proportions of stranger homicides, (e.g., Gibbs, et al., 1977; McClain, 1981; Weisheit, 1984), such a speculation is within the realm of probability. Offense Characteristics (Offender) Premeditation One can see from Table 4 that female homicide offenders for the most part do not pre-plan the murders. Exactly half (50 percent) of the first degree murders reported by Gibbs, et al. (1977) were premeditated; slightly less than half (46 percent) were planned, according to Ward et al. (1969), however the large proportion of unknown cases (38 percent) make this statistic questionable; and in the sub-sample of the totally Black popu- lation examined by McClain (1982), 66 percent of the homicide offenders denied premeditation in the killings. This figure tends to coincide with the second degree (61.1 percent) and third degree murderers (80 percent), who were disproportionately Black in the Gibbs, et al. study (1977), who claimed they did not plan the murder event. Although only three of the studies reported on this factor, they appear to support the notion that women do not plan their homicides in advance. Offender's Role The typology utilized by Ward, et al. (1979: 116) outlines the roles women can play in the commission of violent crimes as: the conspirator, who instigates or has knowledge of the crime but does not participate in committing the criminal act itself; the accessory, who plays a secondary role in committing the crime—acting as lookout, driving a getaway car, carrying weapons, tools, or the proceeds of robberies and burglaries; the partner, who participates equally in all aspects of the crime; and finally, the woman as the sole perpretrator of the crime. (italics in original) I 161 Table 4 l.blr «: Off.nit Ch»r»cl«rl»Mc« (Offrniitr) (In ncrcnnttanO Alcohol Prfwdlt.ted Offtndtr'l >olt InvoWcnent Don't llo Unknown Soli Firlncr Consnjlrttnr Accniiorj Other Offender" THcI 15" Dnlfc f?o.» UnWo.nn IHM) Coir, fl »l. - . •*••■ f7.0 ... 11.0 (0.0 "'•«' 94.0 • - *.0 (7.0 W.rd. rt §1. 71.o' 46.0 3J.0 77.01 lt.0 J.O 3.0 ... i moi Su»»l .ml Hrli;c« - - - 94.4 .... 45.» II174» • fltrnbLtt • nd Greenland - - II.| - - 11.3 - 91.7 Glbbl, et »l. II1IM lit degree 50.0 SO.O 71.4 78.( • - - 71.4 - - *».« 7nd degree 30.9 61.1 - 80.0 70.0 • - - (3.7 - - 17.( )rrf drgrre 70.0 "0.0 - 7S.0 7S.0 - - - SS.3 lot»»n ..... . . . .... ri»;ni "cCUIn - - J4.1 4.1 - 1.7 . 70.0 II1TI) lkCI.1* 33.4 M.( ... . . . ... 77.1 111"?I Wllb.n»» - - - 91.5 ..... 4f., IIWJI I'elihflt - - - 79.4 .... IIVrH) 19(3 an.1 l«M »»n>U». ON to »»t Ion It/ltot We ScH-Oefrnie Alcono1/t>rvg .tuitlf lahle lhl|_lt_tf>l« Influent! SO.* PiyrKolMlr.nl/ Olher' r.uii PiyrKologl Caption*I ntrldrnt/ lrl.lll OiKer/ 17.0/- 78.( II. A H.J 17.( 33.3 Self Preiervatlnn 78.1 JS.3 40.0 Revenge S.91/- (.71/- 11.1 71.3 71.0 11.1 44.4 S9.( 30.0 Hevrnge II.I/- ».» 0.S/7.I 74.0 74.0/- It appears from the eight studies which describe the role of the homicide perpetrator, that a woman who kills, acts alone. The proportions of sole perpetrators, seen in Table 4, range from 71.4 percent (Gibbs, et al., 1977) to 96.4 percent (Suval and Brisson, 1974) with Black female homicide offenders falling on the higher end of the range (94.1 percent) according to the findings reported by McClain (1981). Acting as partners in homicide runs a far second to acting alone in the commission of this crime. Alcohol Involvement Whether the homicide offender, the victim, or both were under the influence of alcohol was noted in a few of the cases. In the present analysis, this category is reported separately from other possible reasons given for the commission of the homicide including whether the offender "blamed" a substance for her deed. These are classified under "rationale/ motive" in Table 4. There is considerable spread across the studies as to whether the homicide offender admitted to alcohol involvement at the time of the crime. The two studies exclusively concerned with Black female homicide offenders suggest very little connection between the use of alcohol and the homicide. Wolfgang (1958), reports only 19.8 percent of the Blacks recalled alcohol involvement, and that both the female offender and her victim allegedly were under the influence in 43.8 percent of the incidents. Similarly, McClain (1981) finds that one-fifth (20 percent) of the Black female homicide offenders were under the influence of alcohol at the time of the homicide. Such findings are not astonishing, since it was noted earlier that female homicide offenders reported a low incidence of alcohol and narcotics abuse. However, in light of this observation, the question is raised whether Black female homicide offenders are less likely than their white counter- parts to introduce the influence of alcohol as an excuse, or rationaliza- tion, for the homicide, since the remainder of the studies indicate fairly extensive alcohol involvement in the crime (from 45.8 percent to 91.7 percent). An examination of the motive given for the homicide helps to partially answer such an inquery. Rationale/Motive for the Homicide Only three studies (Ward, et al., 1969; Gibbs, et al., 1977; and McClain, 1982) found the homicide offender to admit that the influence of alcohol'and/or drugs contributed to the victim's death. The very small proportions (5 percent, 6.7 percent, and 11.1 percent) suggest that contrary to the finding of substance (alcohol or drugs) use on the part of the offender and/or victim during the homicide event, the female homicide I 163 offender did not attribute this usage as a principal cause leading to the violent offense. Although there is considerable scatter across the various reasons proffered for the homicide, the most frequent seems to be for psychological or emotional causation. For the present analysis, this category includes anger and revenge as emotional causes, which appear to be the justification more frequently for Black female homicide offenders, at least according to the McClain (1982) sub-sample in which 44.4 percent listed this reason. Other indicators of this homicide rationale may be inferred from those studies with larger proportions of Black offenders such as Weisheit (1984) who lists revenge as the primary motive in 30 percent of the cases or Gibbs, et al. (1977) who find that second degree (35.3 percent) and third degree murderers (40.0 percent) cited psychological motivations for their crimes. Self-defense, "self-preservation," and other forms of justifiable action were the second most frequent reasons given for killing another, according to seven studies of female homicide. Wolfgang's (1958) Black female sub-group listed self-defense in 50.6 percent of the cases, but only 11.1 percent of McClain's (1982) Black sub-sample cited self-defense as the motive. Claiming innocence may be another way of stating that the homicide was justifiable, but there is no way to verify this from the data reported by Ward, et al. (1969), Suval and Brisson (1974), and Gibbs, et al. (1977). By the same token, with more detailed information, it might be determined that the women who stated the homicide was someone else's fault would also be classified as basically innocent (see Ward, et al., 1969 and Weisheit, 1984). Again, one might presume that the studies which list economic reasons as the rationale for the homicide actually refer to homicides committed during a felony, e.g. robbery; but insufficient information prohibits this supposition. Offense Characteristics Most of the studies examined contained some information on the type of weapon or method of killing used in the female offender homicides, but other factors associated with the crime under consideration here-for example, whether the homicide took place in a residence and the location if outside (bar, street, alley, etc.)- received scanty attention. Table 5 includes the offense characteristics reported. Weapons and Methods With few exceptions, firearms (handguns, shotguns, pistols, rifles, etc.) were the preferred weapons chosen by most female homicide offenders. 164 Biggers (1979: 6) notes that the women murderers in her study (Florida) were familiar with violence either at home or in their neighborhoods, thus "(G)uns were considered 'safe' weapons because attack or defense could be made from a safe distance" and those women who used this type of weapon "either carried guns with them or had easy access to them in their own homes." It has also been pointed out that gunfire is impersonal since the injury may be inflicted from a distance, whereas beating, stabbing and strangulation involve personal contact (Blackbourne, 1984). Poisoning is another homicide method historically believed to be peculiar to women undoubtedly because of the distancing between victim and offender. Interestingly, only two studies (Totman, 1978; Weisheit, 1984) reported the use of poison in the homicides. Totman (1978) indicates that 7.1 percent of the child murders involved poison; not an unusual finding in a parent-child homicide. The other incidence of poisoning, reported by Weisheit (1984) is extremely low, or less than one percent. Another stereotype, that of Black women (and men) as cutters, slashers, and stabbers, may have arisen from findings reported some years earlier by Wolfgang (1958) who reported that Black women tend to use knives, moreso than guns in homicides. As seen in Table 5, the majority of Black female homicide offenders (67.7 percent) in his study did choose knives as weapons. Cole, et al. (1968) also indicate that knives were the most frequent weapon of Black women (48 percent) in homicides. Since 1977, it is apparent that firearms have become preferable to other weapon or method choices for both Black and white women who kill. It is particularly notable that 72.6 percent of the Black women in the six-city study by McClain (1981) and 66.7 percent in the sub-sample (1982) opted for guns in the commission of their homicides. The three studies which clearly identify Blacks separately from white female homicide offenders suggest that Black women are less inclined than white women to use their hands, feet, pipes, ball bats, clubs, and other hand-held clubbing instruments as methods of killing. Wolfgang (1958) finds only 2.2 percent used this means of homicide, Cole, et al. (1968) report a 6.0 percent incidence, or half the white proportion (13 percent), and McClain (1981) also lists only 2.7 percent of Black female homicide offenders selecting hands as the mode of the homicide. Location of the Homicide The paucity of data in the examined studies prohibits a true picture of where a woman elects to murder someone, but the information in Table 5 unmistakably suggests that the site is the residence of either the victim, the offender, or someone else. Wolfgang (1958) found the kitchen to be the most popular homicide location in the home (29.4 percent) which possibly reflects the high incidence of knife homicides he found among the Black female offenders. Twenty years later, however, Totman (1978) reports the bedroom as the most frequent known homicide location (19.6 percent) in mate homicides; possibly this is where the gun is kept. 165 TABLE 5: OFFENSE CHARACTERISTICS TYPE OF WEAPON/METHOD USED FIREARM KNIFE HANDS1 POISON OTHER/UNKNOWN WOLFGANG (1958) 22.5 67.7 2.2 COLE, ET.AL. (1968) White 37.0 32.0 13.0 Black 31.0 48.0 6.0 DUARD, ET AL. (1969) ROSENBLATT And GREENLAND (1974) 34.0 35.0 8.0 50.0 - 50.0 GIBBS, ET AL. (1977) 1st deqree 100.0 2nd degree 73.7 15.8 3rd degree 62.5 31.3 TOTMAN (1978) Mate Child MCCLAIN (1901) MCCLAIN (1982) WILBANKS (1903) WEISHERT (1984) 61.0 25.0 - 71.5 7.1 72.6 23.0 2.7 66.7 11.1 11.1 59.6 25.5 14.9 44.0 40.0 ? 0.007 7.5 18.0 15.0 23.0 10.6 6.3 14.0 21.5 1.7 11.1 (in percentages) HOMICIDE LOCATED IN RESIDENCE OF: AWAY FROM LOCATION IN RESIDENCE LOCATION OUTSIDE RESIDENCE LIVING BED STAIRS OFFENSE VICTIM BOTH OTHER RESIDENCE ROOM ROOM KITCHEN HALL OTHER BAR STREET OTHER/UNICIOWfl 6.0' 64 .0 47.0 2.0 47.0 45.0 59.0 14.0 22.0 15.4 37.6 25.6 In the home (not Identified) 63.0 9.2 25.7 29.4 9.2 3.7 12.0 10.1/ 13.7 19.6 5.9 23.5 5.9 5.9 13.7/11.8 2.6 29.0 6.0/ - 2.1 10.6 12.8/ 1. INCLUDES BEATING, CLUBBING, STANGLING 2. 1963 and 1960 Sample Finally, very few female homicide offenders commit their offenses away from the residence, i.e. on the street, in an alley, parking lot, automobile, or a bar. It is notable that McClain (1981) reports the street as the homicide locale in 29 percent of the Black female homicide cases, a figure almost three times that found by Wilbanks (1983b) and, according to Totman (1978), not quite five times the frequency among mate killings. Criminal Justice Experience The final data, seen in Table 6, describe the criminal justice system experience of the female homicide offender. The most abundant findings concern previous arrests and are reported by nine of the research efforts. Unfortunately, the fate of the offender after apprehension appears to have been overlooked, ignored, or was unavailable to most of the researchers. Previous Arrests A curious pattern differentiating white and Black female homicide offenders appears after a careful analysis of each of the studies reporting previous arrest histories of the offenders. Research including only Blacks and those instances in which Blacks can be examined apart from whites seem to indicate that Black female homicide offenders tend to have been involved in the criminal justice system more frequently than their white counterparts. Wolfgang (1958), for example, shows that slightly more than half (51.6 percent) of the Black female homicide offenders had previous arrests. This finding is also noted by McClain (1982) who found 62.5 percent in this category, Cole, et al. (1968) who report 81 percent, and Totman (1978), who lists 90 percent Black female mate killers with previous arrests. Although Totman found lower proportions of child killers than mate killers with previous criminal records, nonetheless the Black women who killed their children (33.3 percent) were three times more likely to have previous arrests than white women committing the same offense (10 percent). Other indications of a difference between white and Black women on this variable may be seen in the lower percentages of previous arrests among white female homicide offenders in studies where they predominate. Rosenblatt and Greenland (1974), for instance, report on Canadian women, whom one would presume are mostly white, and only 16.7 percent were found to have been arrested prior to their homicide arrests. In the Ward, et al. (1969) study, 64.3 percent of the group revealed previous arrests. On the other hand, Suval and Brisson (1974) report a sample with 80.5 percent Black, yet the previous arrest history of the group is only 27.1 percent. Also, Totman (1978) records 65 percent of the white mate killers with prior arrests. Obviously no definitive statement may be made about the phenomenon, but there is the suggestion that Black female homicide offenders tend to have had more previous contact with the criminal justice system. This is not surprising since minority women, 167 Table 6: Criminal Justice Offender Information (in percentages) 00 Previous Arrests YES NO Wolfganq (1S58) 51.6 40.4 Cole, et al. (1968) White Black 66.0 81.0 33.0 19.0 Ward, et al. (1969) 20. Q2 BO.O Suval and Brisson (1974) 27.1 72.9 Rosenblatt and Greenland (1974) 16.7 83.3 Gibbs, et al. (1977) 1st Degree 2nd Degree 3rd Degree - - Totman (1978) Mate: Black White Child: Black White 90.0 65.0 33.3 10.0 10.0 35.0 66.7 90.0 McClaln (1902) 62.5 37.5 Court Disposition GUILTY DIsliiTsTed/ ACQUITTED 66.3 27.2 OTHER 6.5 Pr 1 son Sentence (Years) TT----STB 9^T5 15-30 30-L1fe Life Death None 25.0 25.0 25.0 25.0 12.5 - - 14.3 85.7 5.0 20.0 35.0 40.0 37.5 43.8 6.3 - S» a »—< C7J Wilbanks (1983) 63*0 18.5 18.5 41.2' 17.6J 17.6 11.8 11968 Sample. 9 Up to 5 years, 6-10 years. 'll-lO years. similar to minority men, find themselves more frequently entangled in the law enforcement net for a number of reasons: more surveillance by the police in minority communities, police prejudice against minorities, wide police arrest discretion, and the many social ills and conditions in the social structure which contribute to deviance that are more frequently located in poor, and minority neighborhoods. Court Disposition and Sentencing Only two studies report on the dispositions of the courts in female homicide cases. Wolfgang (1958) found that 66.3 percent of the Black female homicide offenders were found guilty, while 27.2 percent were dismissed or acquitted. More recently, Wilbanks (1983b) reports that 63 percent of the women were found guilty and 18.5 percent dismissed or acquitted in a sample that consisted of 59.6 percent Black and 40.4 percent white female homicide offenders. If these two studies may be used as guideposts, it would seem that little change in the court dispositions of female homicide cases has taken place in 25 years. Women who commit homicide do not appear to receive excessively long prison sentences. As seen in Table 6, only the Florida study reported by Gibbs, et al. (1977) seems to suggest harsh sanctions where 14.3 percent of the first degree murderers (85.7 percent of whom are white) were sentenced to 15 to 30 years in prison and 85.7 percented received prison terms of 30 years to life. Yet it is curious that the proportion of second degree murderers (35 percent) in the Gibbs, et al. sample (which is 65 percent Black) is twice that of the first degree murderers for sentences of 15 to 30 years. Another unusual finding on this factor is seen in Wilbanks' (1983b) sentencing percentages for Dade County (Miami) Florida. It appears that female homicide offenders in Miami were not treated as harshly as the second and third degree female murderers in Florida prison reported by Gibbs, et al. (1977). Only 17.6 percent of the Miami offenders received six to ten years for homicide, while the same proportion were sentenced to 11 to 20 years, and a surprising 41.2 percent only received up to five years in prison. Since Wilbanks did not include the degree of murder, it is possible that the female homicide offenders he describes would come under the second or third degree categories. In the studies examined, none of the women who committed the offense of homicide received a death sentence. However, on November 2, 1984, the first woman to be executed in the United States since 1962 was killed by lethal injection in North Carolina. As of October 1, 1985, there were 19 women (1.19 percent), among the 1,590 individuals, on death rows in various states of this country (NAACP, 1985). More than one-third (36.8 percent) of these female criminal homicide offenders are minority women, six Black and one Native American. An earlier examination of the eleven women on death row on May 8, 1982 revealed that one of the two Blacks had poisoned her abusive husband and the other vis waiting in a car when her husband killed a white man (Mann, 1984). ** jng the murders committed by 169 the white female offenders at that time on death row: two poisoned for money, one masterminded the murder of her former husband for insurance money, which was carried out by her present husband, one was a "hit" murder for money, two were rape-murders involving multiple stabbing (97 times) in one case and strangulation in another, one shot her lover's wife in an insurance conspiracy, another kidnapped and "executed" five of her boyfriend's former in-laws, including a two-year old child, and one with the help of a male accomplice slowly strangled two men, one of whom had given them a ride and a place to spend the night. From these brief descriptions, it seems that the white female criminal homicide offenders, for the most part, were more culpable, and the homicides were far more heinous than those of the Black female offenders. If these homicides cases are indicative of the current female death row occupants, it suggests there may be differential sentencing of women because of racial status. The paucity of data on the criminal justice processing of women arrested for homicide and the scarcity of other pertinent offender and offense characteristics clearly emphasize the glaring need for more research on this important topic. Analysis of City Data As previously discussed, a survey instrument was designed and submit- ted for completion to previously contacted top-level law enforcement administrators in Atlanta, Chicago, New York City and Houston. Through a detailed examination of the Uniform Crime Reports, each of these cities had been identified as having high murder rates. Further, it was hoped that these four cities, located geographically as they are, would provide a modest regional picture of the phenomenon. It is clear from Table 7 that the in-depth and detailed data petitioned for were quite ambitious considering the numbers involved forwarded by the cities. New York City (McGuire, 1977) cleared 79 murder cases with female offenders in 1977, Atlanta, Chicago (Brzeczek, 1979), and Houston in 1979 cleared 29, 99, and 89, respectively; while in 1983, Chicago recorded 81 such cases (Rice, 1983), Houston, 77 (Brown, 1983), and Atlanta, 23 (Napper, 1985). These numbers, total 477 female murderers, so it is readily understood why the four cities contacted were unable to complete the survey instrument requested on each offender. The Findings Race Regardless of the city or year, Black women are obviously the predom- 170 inantly arrested female offenders in cleared murder cases. As Table 7 depicts, these offenders represent from 61 percent (Houston, 1983) to 87 percent (Atlanta, 1983) of the women arrested for murder. The finding that Black women are disproportionately arrested for murder in Atlanta, Chicago, New York City and Houston coincides with the results of the studies reviewed above. Even more interesting are the proportions represented by Black women of the total murder arrestees in these cities. Table 7 indicates that Black women range from a low of 5.7 percent of New York's murderers in 1977, to 14.2 percent in Atlanta in 1983, whereas the range for white female murderers is less than one percent for New York City and at the highest, 5.2 percent in Atlanta. In other words, white females' highest proportion of total murder arrests is less than the lowest proportion for Black females. According to the Uniform Crime Reports, in 1979 females comprised 13.7 percent of arrests for murder and nonnegligent manslaughter (UCR, 1980: 199). The city of Atlanta in 1979 records 18.7 percent of murder arrests as females, the highest of any of the cities included here, and of this number 13.5 percent are black. Although Chicago lists a little less than the national female average for that year in murder arrests (12.3 percent), Black women again make up the majority at 10.4 percent. At the national level, in 1983 there was a slight decrease (13.3 percent) compared to 1979 in the proportion of females arrested for murder and nonnegligent manslaughter (UCR, 1984: 186). That year Chicago also experienced a decrease (11.1 percent), but Black females were nine times the percentage of white females as a proportion of total arrests for this offense category. The data available for Houston indicate that this city in 1983 had a very high proportion of females arrested for murder (16.4 percent) compared to the national figure for that year, with Black females at over twice the percentage (10 percent) of white females (4.5 percent). A Regional Observation Rather interesting differences are noted between the cities in terms of the incidence of female homicide and their geographic locations in the United States. Regionally, New York was selected to represent the east; Chicago, the midwest; Atlanta, the south; and Houston, the west. It is debatable whether the city of Houston is considered as western or southern, therefore we may consider it both. Mindful of the small sample and the differences in years, it is still interesting to note that there seems to be an increase in the proportion of females arrested for murder as one goes from east to west and north to south. Among all arrests for murder, New York City (1977) has the smallest percentage of females (8.7 percent), followed by Chicago in both 1979 (12.3 percent) and 1983 (11.1 percent). On the other hand, the two southern cities, Atlanta and Houston, reveal much larger female involvement in murder, according to their arrest statistics. Atlanta, in 1979, for 171 Table 7 Cit^ City Study - Cleared Murder Cases With Female Offenders By Race Totals Black White Latin (Year) N ( New York 79 ( (1977) (% of all murders) (n=910) ( Atlanta 29 ( (1979) {% of all murders) (n=155) ( Atlanta 23 ( (1983) {% of all murders) (n=141) ( Chicaqo 99 ( (1979) (% of all murders) (n=808) ( Chicaqo 81 ( (1983) (% of all murders) (n=729) ( Houston 89 ( (1979) (% of all murders) (n=702) ( Houston 77 ( (1983) (% of all murders) (n=471) ( %) N~ ~~(%) N (%) N (%) 100.0) 52 (65.8) 8 (10.1) 19 (24.1) 8.7) 18.7) 12.3) 11.1) 12.7) 16.4) (5.7) (0.9) 100.0) 21 (72.4) 8 (27.6) (13.5) 100.0) 20 (87,0) (5.2) (13.0) (10.4) (1.6) 100.0) 68 (84.0) (9,3) (1,0) (8,8) (2,6) (10.0) (4.5) (2.1) 16.3) - (14.2) - (2.1) - 100.0) 84 (84.9) 13 (13.9) 2 (2.0) (0.3) (8,6) 5 (6,2) (0,7) 100.0) 62 (69.7) 18 (20,2) 9 (10.1) (1,3) 100.0) 47 (61,0) 21 (27,3) 9 (11.7) (1.9) Other N (%) (1.2) (0.1) 1 1979 and 1983 data not available 2 These data derived from previous Atlanta research by the author 172 example, had fifty percent more females represented among those arrested for murder (18.7 percent) than Chicago the same year (12.3 percent) and twice the female proportion for murder that New York had in 1977 (8.7 percent). Houston, in 1983, reveals similar figures, since that year 16.4 percent of all persons arrested for murder were females, a little less than fifty percent more than the proportion for Chicago in 1983 (11.1 percent). If Houston is defined as a western city, it would still be determined to differ from the north and east in having a higher percentage of arrested female murderers, but would trail 1979 Atlanta in this category by 2.3 percentage points. If there is any possible validity to such regional differences, one might speculate that in "old" cities like New York and Chicago which are more urbane, women may be more equal to men, less frustrated and thus less prone to murder. Whereas in rapidly developing "new" cities, such as Atlanta and Houston, women may come under more pressure than women in "settled" urban areas, experience a less egalitarian position in the social structure, and consequently explode in violent behavior. The vagaries of the southern culture could be contributory to female violence, as could the "western frontier" tradition. Summary and Conclusions Black female homicide offenders have been demonstrated through arrest statistics to rank second, next to their Black brothers in the incidence of taking another human life. Part of this problem may be attributed to institutional racism in the criminal justice system that leads to the processing of non-whites more extensively than Blacks for the commission of offenses. The deplorable conditions in minority communities, e.g. poor housing, lack of health care, underemployment, unemployment, racial discrimination, and many other social ills in the social structure no doubt contribute substantially to the minority crime problem. None the less, the findings of the current study suggest that the disproportionate involvement of Black females in homicide can not be totally explained by either institutional racism or defects in the social structure. The fact that Black women comprise only about eleven percent of the female population in the Unites States and yet are arrested for almost three-forths of the murders committed by females leaves an explanatory gap that demands to be closed. Homicide among Blacks in this country is believed to have reached epidemic proportions. Since homicide is an intraracial event, it is obvious that Blacks are slowly eliminating each other almost to the point that one could speak of Black-on-Black "genocide." In addition to the decimation of each other, Blacks accused and convicted of murder when imprisoned add to the growing number of minorities in the correctional systems of this country that are far out of proportion to their numbers in the population. This paper cannot answer why this is happening, but by offering a profile of the Black female homicide offender, it is hoped that 173 this will be a first step in addressing the problem of Black violence. This paper includes information gleaned through an exhaustive search of the literature on female homicide offenders which was supplemented by more recent arrest data obtained from urban areas that contain large numbers of Blacks and are believed to represent a partial regional sampling of the country: New York City (east), Atlanta (south), Chicago (midwest), and Houston (west and southwest). A total of fifteen empirical studies focusing on female homicide offenders over a twenty-six year period (1958-1984) were identified and form the basis of the analysis. These studies, three of which included identifiable Black subjects, were examined, and where possible the Black female offenders were isolated in the others. Throughout, however, it was felt to be invaluable to maintain the white homicide offender data for comparative purposes. As previously indicated, Black females tend to predominate among homicide offenders with the finding that they comprise about 73.6 percent of the female arrests for murder, whereas white women are only an average of 17.5 percent of such offenders. Both groups tend to be older than the typical female offender in the general prison population. Most incarcer- ated women are under thirty years of age, but the woman who kills, Black or white, tends to be over thirty. Similar to other incarcerated v/omen, female homicide offenders are reported to have lower intelligence (as measured by Intelligence Quotients) and concomitantly, lower education attainment. Black female homicide offenders were found to have lower I.Q.s and less years of completed education that white women who committed the same offense. It was further noted that the I.Q. of the female homicide offenders varied inversely with the complexity of the homicide, a finding first noted by Figueria-McDonough (1981). With little education and less than normal intelligence, it is not surprising to find that females who commit homicide are more likely, when employed, to be located in occupations in the lower status levels, such as service jobs. This is more obvious among Black female homicide offenders. Previous studies of incarcerated female offenders have shown that our prisons are mostly occupied by single women, with Black women more frequent- ly occupying an unmarried status than whites and other minority groups. In the present research examination, an opposite tendency was found, in that female homicide offenders, regardless of race, are more likely to have been married or at least cohabiting with men. It was noted that since most of the studies were undertaken in southern states, marital status may be indicative of the mores and moral standards more peculiar to the south. Another social characteristic frequently ascribed to female offenders is that they are products of broken homes. Overall, this was not found to 174 be true of female homicide offenders; again, irrespective of racial status. One of the stereotypes of females who commit crimes is that they are "fallen women." Many references have been made to the "madonna/whore" dichotomy which suggests that there are "good" women and "bad" women. The good women adhere to the double standard inherent in this county by being housewives and mothers. On the other hand, the bad women are the whores, the sexually promiscuous or deviant, the criminals. The research efforts examined in this paper belie the myth of the fallen woman by finding little reported incidence of sexual promiscuity or homosexual activity among the female homicide offenders. Although only reported by two of the studies, Black women were found to deny sexual promiscuity and homosexuality more frequently than the white women questioned about such behaviors. It is often felt that in the commission of murder, the offender and possibly the victim, as well, are under the influence of alcohol or drugs. This investigation found little support for such a contention among the women examined in the studies available. Not only was there little indication of alcohol or narcotics use among female homicide offenders, but it was also unlikely that these offenders would use substance abuse as an excuse for the homicide, even though a small proportion of the studies indicated that they were under the influence at the time of the crime. In line with other studies of homicide, it was clearly found that homicide is intersexual, intraracial, and intrafamilial. Female homicide offenders, and Blacks are no exception, tend to kill their mates, or those members of their families closest to them. Black women do tend to kill friends and acquaintances more than their white counterparts, and there is some indication that these victims may be other females. Women, as a rule, are infrequently involved in stranger homicides, which typically result from the commission of other crimes, such as robbery or burglary. Women also do not tend to plan their murders. Black female homicide offenders are even more unlikely to premeditate their homicides than white females incarcerated for this crime, according to the studies reporting on this topic. Further, unlike other offenses, e.g. burglary or robbery, Black female offenders who kill are more likely to be the sole prepetrator, and not accessories or partners to the event, than white female homicide offenders. It is suggested by the data examined that the reason for such a finding is apparently related to the rationale or motive given for the homicide. Psychological or emotional reasons, typically anger and revenge, are most frequently given by Black v/omen for committing murder. Data from New York, Chicago and Houston corroborate this finding by indicating that arguments or altercations are the most frequent apparent motive indicated in murder. Among the female homicide offenders examined, self-defense is the second most frequent reason given for the homicide, which further indicates some degree of emotionality and an altercation between the victim and offender that led to the event. Women, and again, Blacks are no exception, tend to kill their victims in the home through the use of firearms. Earlier studies had suggested 175 that Black female homicide offenders were more likely to use knives or other cutting instruments as murder weapons, but this choice of method, over time, has been replaced by guns which are considered "safe" and impersonal weapon choices because of the distance they put between the prepetrator and the victim. The present analysis indicates that Black women who murder have been previously involved in the criminal justice system more frequently than white women who kill. The indicator for this finding is seen in the previous arrest history of the offenders, but the differing results across the studies reporting such data renders this suggestion somewhat specula- tive. Even if verification could be documented, it would not be too surprising since Black people, in general, are frequently the victims of police surveillance and harassment which results in more arrests, and ultimately, sentencing. On this last point, there are clues from a few of the studies, that once convicted, Black female homicide offenders are more likely to receive harsher sentences compared to white females who commit the same crime. The controversy over race and sentencing is still an integral part of the criminological dialogue, but it is suggested here that Black women of- fenders, similar to their Black brothers, are treated differently by the criminal justice system. It is obvious that a great deal of research is still needed on this topic before any definitive statements can be made about the Black female homicide offender. There is a paucity of studies on this topic and those available vary in research focus and concomitant results. This paper has attempted to put together a profile of the Black female who murders, through a minute examination of every study available on this offender group. Unfortunately, the dearth of such studies prohibits a final portrait of the Black female homicide offender, but at least, a snapshot of her has been provided. 176 FOCrrNOfTES Actually this figure is reported for only the female homicide inmate population in California in 1968. The Ward, et al. data are difficult to decipher, since they combine some of the samples on some charac- teristics, report separately for others, and on occasion collapse the black offenders into "nonwhite" categories. The female homicide offender profile reported in "Crimes of Violence by Women, " (1969) is thus a composite of several samples. Rather than try to reconstruct their data, the tables and text in this paper retain the information as Ward, et al. presented it and note the sample sources by year. In cases where the data are given for separate years, 1968, the latest year, is reported. This consequential study will be referred to throughout the paper as a valuable source of basic social and demographic data on women offenders in the United States. Initially this invaluable research was seriously considered for inclusion in the paper because of its wealth of social data on incarcerated female homicide offenders in Michigan but since the study was not exclusively concerned with the offender group under examination it was decided to include the pertinent homicide informa- tion throughout the body of the text when relevant. According to information obtained from the NAACP Legal Defense Fund, these women are in Alabama (one Black, one white), Arkansas (one white), Florida (one Black), Georgia (one Black, one white) Idaho (one white), Indiana (one white), Maryland (one white, one Native American), Mississippi (one white), Nevada (one Black, one white), New Jersey (one white), Ohio (two Blacks, one white), Oklahoma (one white) and Texas (two whites). Murder is the term used by law enforcement and the Uniform Crime Reports and will be used regarding the city data. Research in progress by the author includes random samples from homicide files on female offenders from Chicago and Houston. These, and data from other urban areas, will be analyzed and reported in a later paper. 177 REFERENCES ADLER, F. (1979) "The Interaction Between Women's Bnancipation and Female Criminality: A Cross-Cultural Perspective." In F. Adler and R. Simon (eds.) The Criminology of Deviant Women. Boston: Houghton Mifflin Company. BIGGERS, T.A. (1979) "Death by Murder: A Study of Women Murderers." Death Education 3: 1-9. BLACKBOURNE, B.D. (1984) "Women Victims of Homicidal Violence." Unpublish- ed paper. BROW, L. (1983) Annual Report 1983 Homicide Division. Houston, Texas. Xerox copy. BRZECZEK, R.J. (1979) Murder Analysis 1979. Chicago: Chicago Police Department. OOLE, K.E., G. FISHER and S.R. COLE (1968) "Women Who Kill." Archives of General Psychiatry 19: 1-8. DEMING, R. (1977) "The Black Female Criminal." In Women: The New Criminals. New York: Thomas Nelson and Sons. EBBE, O. (1984) "The Correlates of Female Criminality in Nigeria." Unpublished paper. FIGUEIRA-MCDONOUCH, J. (1981) "Profiles of Female Offenders." In J. Figueira-McDonough, A. Iglehart, R. Sarri, and T. Williams (eds.) Females in Prison in Michigan 1968-1978: A Study of Commitment Patterns" Ann Arbor, Michigan: School of Social Work and the Institute for Social Research. GECRGES-ABEYIE, D. (1984) The Criminal Justice System and Blacks. New York: Clark Boardman Company, Ltd. GIBBS, D.L., I.J. SILVERMAN and M. VEGA (1977) "Homicides Committed by Females in the State of Florida." Paper presented at the annual meeting of the American Society of Criminology. GLICK, R.M. and V.V. NETO (1977) National Study of Women's Correctional Programs. U.S. National Institute of Law Enforcement and Criminal Justice. Washington, D.C: U.S. Government Printing Office. HINDELANG, M.J. (1974) "The Uniform Crime Reports Revisited." Journal of Criminal Justice 2: 1-17. 178 IGLEHART, A. (1981) "Personal and Social Characteristics of Female Offenders." In J. Figueira-McDonough, A. Iglehart, R. Sarri, and T. Williams (eds.) Females in Prison in Michigan 1968-78: A Study of Commitment Patterns. Ann Arbor, Michigan: School of Social work and the Institute for Social Research. JASON, J., L.T. STRAUSS and C.W. TYLER (1983) "A Comparison of Primary and Secondary Homicides in the U.S." American Journal of Epidemiology 117 (3): 309-319. JONES, A. (1980) Women Who Kill. New York: Holt, Rinehart and Winston. KLECK, G. (1981)) "Racial Discrimination in Criminal Sentencing." American Sociological Review 46: 783-805. LEWIS D.K. (1981) "Black Women Of fenders and Criminal Just ice." In M. Warren (ed.) Comparing Female and Male Offenders. Beverly Hills, CA: Sage Publications. MANN C.R. (1981) "The Minority Woman Offender in the Criminal Justice 'system." Unpublished paper presented at the annual meeting of the American Society of Criminology. MANN, C.R. (1984) Female Crime and Delinquency. University, Ala.: University of Alabama Press. MERCY, J.A., J.C. SMITH and M.L. ROSENBERG (1983) "Homicide Among Young Black Males: A Descriptive Statement." Paper presented at the annual meeting of the American Society of Criminology. MESSNER S.F. (1983) "Sex Differences in the Societal Arrest Rate: A ^ ^ss^tional Test of the Theory of Structural Strain." Paper presented at the annual meeting of the American Society of Criminology. MCCLAIN P D. (1982a) "Cause of Death-Homicide: A Research Note on Black Females as Homicide Victims." Victimology 7:204-212. rurriAMi P D (1982b) "Black Female Homicide Offenders and Victims: Are "^ey From the Same Population?" Death Education 6: 265-278. MCGUIRE, ROBERT J. (1977) Homicide Analysis. New York: Office of Management Analysis, Crime Analysis Section. NAACP (1985) "Death Row, U.S.A." New York: NAACP Legal Defense Fund. Mimeograph. NAPPER, G. Information received in 1985 from Commissioner Napper, Atlanta. POPE C.E. andR.L. HCNEELY (eds.) (1981) Race, Crime and Criminal 'justice. Beverly Hills, CA.: Sage Publications. 179 RASKD, G. (1976) "The Victim of the Female Killer." Victimology 1:396-402. RICE, F., Jr. (1983) Murder Analysis 1983. Chicago: Chicago Police Department. RIEDEL, M. (1984) "Blacks and Homicide" In G. Daniel The Criminal Justice System and Blacks. New York: Clark Boardman Company, Ltd. RIEDEL, M. and L. LOCKHART-RIEDEL (1984) "Issues in the Study of Black Homicide." Paper presented at the annual meeting of the American Society of Criminology. ROSE, H.M. and P. MCCLAIN (1981) Black Homicide and the Urban Environment. Washington, D.C: Center for Studies of Minority Group Mental Health, National Institute of Mental Health. ROSENBLATT, A. and C GREENLAND (1974) "Female Crimes of Violence." Canadian Journal of Criminology and Corrections. 16: 173-180. SHIN, L., D. JEDLKXA and E.S. LEE (1977) "Homicide Among Blacks." Phylon 398-407. SKDGAN, W.G. (1974) "The Validity of Official Crime Statistics: An Etapirical Investigation." Social Science Quarterly 55: 25-38. STAPLES, R. (1975) "White Racism, Black Crime, and American Justice: An Application of the Colonial Model to Explain Crime and Race." Phylon 36: 14-22. SUFRIN, M. (1984) "Female Felons, Lady Killers, and Other Women in Crime." Genesis November, 1984: 25-27. SUVAL, E.M. and R.C BRISSON (1974) "Neither Beauty Nor Beast: Female Criminal Homicide Offenders." International Journal of Criminology and Penology 2: 23-34. SUniERLAND, E.H. and D.R. CRESSEY (1978) Criminology, 10th edition. Philadelphia: J.D. Lippincott. TOIMAN, J. (1978) The Murderers: A Psychosocial Study of Criminal Homicide. San Francisco: R. and E. Research Associates. U.S. DEPARTMENT OF JUSTICE (1980) Uniform Crime Reports for the United States. Washington, D.C: U.S. Government Printing Office. U.S. DEPARTMENT OF JUSTICE (1984) Uniform Crime Reports for the United States. Washington, D.C: U.S. Government Printing Office. 180 WARD, D.A., M. JACKSON and R.E. Ward (1969) "Crimes of Violence by Women." In D. Mulvihill and M. Tomin (eds.) Crimes of Violence. National Commission on the Causes and Prevention of Violence, staff report 11-13. Washington, D.C: U.S. Government Printing Office. WEISHEIT, R.A. (1984) "Female Homicide Offenders: Trends Over Time in an Institutionalized Population." Justice Quarterly 1 (4): 471-489. WILBANKS, W. (1982a) "Murdered Women and Women Who Murder: A Critique of the Literature." In N.H. Ratter and E.A. Stanko (eds.) Judge, Lawyer, Victim, Thief: Women, Gender Roles, and Criminal Justice. Boston: Northeastern University Press. WILBANKS, W. (1983a) "Female Homicide Offenders in the U.S." International Journal of Women's Studies 6 (4):302-310. WILBANKS, W. (1983b) "The Female Offender in Dade County, Florida." Criminal Justice Review 8 (2): 9-14. WILBANKS, W. (1984) Murder in Miami. New York: University Press of America. WILBANKS, W. (1985a) "Is Violent Crime Intraracial?" Crime and Delin- quency, (forthcoming). WILBANKS, W. (1985b) "Criminal Homicide Offenders in the U.S.: Black vs. White." In R. Davis and D. Harokin (eds.) Black Suicide and Homicide (forthcoming). WOLFGANG, M.E. (1958) Patterns in Criminal Homicide. Philadelphia: University of Pennsylvania Press. WYRICK, E.S. and O.H. Owens (1977) "Black Women: Income and Incarcera- tion." In C Owens and J. Bell (eds.) Blacks and Criminal Justice. Lexington, Massachusetts: Lexington Books. YOUNG, V. (1980) "Women, Race, and Crime." Criminology 18: 26-34. 181 Can We Substantially Lower Homicide Risk in the Nation's Larger Black Communities? Harold M. Rose Professor Department of Geography and Urban Affairs University of Wisconsin at Milwaukee Milwaukee, Wisconsin Can We Substantially Lower Homicide Risk in the Nation's Larger Black Communities? Introduction Measured both in terms of levels of risk and percentage of people dying from a specific cause, homicide as a cause of death has increased in importance over the last two decades. Changing levels of risk in association with homicide deaths during this interval have received much attention from scholars representing various disciplines (Barnett, Kleitman and Larson, 1975; Blau and Blau, 1982; Farley, 1980; and Holinger and Klemen, 1982). But most of this work tends to focus on causal factors responsible for the increase or the testing of scholarly theories associated with the behavioral propensities of individuals and groups. Much of our understanding of changes occurring in homicidal behavior is the result of national aggregate analyses or individual case studies conducted at varying time intervals. It should be noted, however, that some of the latter studies employ longitudinal data and make a real contribution to our understanding of changes in risk over time and from one environment to another (Block, 1975; Block and Zimring, 1973). Nevertheless, only a few of these studies have focused exclusively on the subpopulation at highest risk—black Americans—or how we might effectively reduce risk. Therefore, the goal of this essay is to detail the status of homicide risk in the nation's larger black communities and to offer some tentative steps designed to lower risk to more acceptable levels. The task identified above is not an easy one as the data required for such an undertaking are not always readily available. Moreover, the extant scholarly literature provides no clear direction as to how one might go about this task. The former shortcoming is associated with the past tendency of the two primary data collection agencies that compile homicide statistics (the FBI and the National Center for Health Statistics) not to publish subaggregate statistics, e.g., individual urban places where racial headings are employed. Therefore, until recently, racial differences in both level of victimization and risk of victimization were only readily available at the national level. At the local level, the situation is mixed, with investigators often choosing to examine the problem in places where police deparrtments and local health departments cooperate more easily. So one of the reasons for our lack of understanding of local differences in risk based on race is directly related to the absence of data at this level. A second—but somewhat less critical—element is the absence of a body of scholars who have a commitment to assist in providing answers to the continuing problem of high homicide risk in the nation's larger black communities. Because there is a lack of ongoing interest, we have a 185 series of single case studies prepared by individuals whose primary research interest lies elsewhere. Thus there is a major void in the literature that represents a continuation of growth and development around aspects of the problem. No doubt the foregoing partially reflects the criminal orientation of much of the research on the topic, and thus a hesitancy to employ race as a focusing variable is somewhat understandable. Because of this tendency, a public health approach to the problem is likely to prove less threatening and at the same time is likely to encourage continuous investigation of the topic. Until such time as the foregoing impediment is removed, the task of developing and/or recommending a robust set of procedures designed to alter risk levels in the nation's larger black communities will represent an extremely complicated and tentative task. Yet when one considers the seriousness of the problems, steps--however tentative--must be developed, evaluated, and acted upon if deemed meritorious. Of course, the issues of data availability and research focus are crucial to our being able to effectively assess the seriousness of the situation. Nevertheless, those issues will no doubt diminish in importance over time such that we can more readily evaluate recent changes in risk. Our primary focus will be to look closely at patterns of risk that evolved during the 1970's and to attempt to extract insights from these patterns. These insights will be helpful in attempting to lower risk to more reasonable levels, given the position of the United States in the world economy. That is to say, homicidal risk in the nation's larger black communities more closely resembles the high levels of risk in developing countries than the lower levels of risk in developed countries (Day, 1984). Yet we must be careful not to establish goals for which there is little chance of success in achieving. Given the circumstances that many black individuals find themselves in and the quality of life in their communities, the task of specifying an optimal level of risk is one that we might wish to ignore. But to do so would simply indicate that we have little hope in the future and that we might simply wish to view homicide vic/timizations as random acts over which we have no control. Such a defeatist posture should be discouraged; and based on current reality, a set of goals should be established that we are willing to work hard to achieve. Establishing the Target Goal The level of optimality that seems minimally reasonable—and the one suggested here—is 25 homicides per 100,000 black persons in the population. The suggested level is two and one-half times higher than that which characterized the white population in 1980. Since the suggested level of optimality was previously attained, this level does not appear to be unreasonable. In 1960, the homicide rate recorded for non-whites was estimated to be approximately 22 per 100,000 (Shin, 1981). 186 The rate hovered around that level for almost a decade prior to experiencing a sudden takeoff after 1965. Therefore, it is possible to constrain acts of lethal violence to levels that might be viewed as more acceptable. The context in the 1980's obviously differs from that prevailing in the 1950's when homicide levels were decreasing. Recently, however, evidence was presented that showed homicide rates for non-whites were already showing downward movement (Ueshima and others, 1984). A corresponding increase in rates of white victimization during the same interval has led to a reduction in the ratio of black to white homicide deaths. By 1980, the gap in the racial differential in homicide victimization had declined to a level of 6:1. During the 1930's, the gap reached a high of 11:1. Thus one of the primary goals of any effort to prevent homicide should be to further lower the victimization gap between the races. In order to attain the established goal, it will be necessary to lower the rate prevailing in 1980 by approximately one-third. If the effort is successful, this would lead to an annual reduction in the loss of lives of the order of magnitude of 2,700 individuals based on the 1980 population. It is unlikely, however, that such a reduction can be achieved in a brief time interval. Neverthless, if serious effort were devoted to the enterprise, we would suggest that this goal could be attained by 1990. The first step toward facilitating the achievement of this goal would be to identify those segments of the population in which risk is most attenuated and attempt to establish the causal factors that promote extreme vulnerability. The overall goal of any preventative effort would be to lower risk across all subgroups in the population. It should be acknowledged, however, that the task will vary in terms of ease or difficulty as a function of the strength of the causal factors. Vulnerable Populations Elevation in homicide risk within the black population has been characterized by a disproportionate heightening of risk among specific segments of that population. During the most recent upturn in violence, the most vulnerable segment of the black population was young adult males (20-24 years old). Between 1960 and 1974, the risk level in this age group increased 105%. A similar but slightly smaller increase in risk characterized the 15-19 year old male population during the interval. Although risk in specific age groups was more attenuated than in others, no five-year age group was left untouched by significant increases in risk during this period of social upheaval. Although the more notable increases in risk have involved black males, the impact of elevated risk has not been confined to men only. The rate of increase in risk for black females over this period was just under one-half that characterizing black males. The ratio of black male to 187 black female risk is approximately 4:1, yet the latter group is at higher risk than white males. Changes in risk for black females were most notable among those 20-24 and those 70 and over. At other ages, changes were less dramatic. It should be noted, however, that even though rates of risk for black females are much lower than those of black males they manifest unusually high rates for their gender. Thus in attempting to identify vulnerable populations who were responsible for the largest share of increasing risk over a 15-year period, we find that young adult males (15-34) and young adult females (15-24) and older females (> 70) were the most vulnerable. Therefore, prevention programs designed to reduce risk to approximately the I960 level should look carefully at these groups for clues that might aid in reducing risk. Target Environments The recent epidemic of homicide has essentially been confined to the nation's larger urban centers and especially to those in which a) there was a large black base population (>250,000) at the time of risk takeoff; b) in which the black population was growing rapidly; c) and in which economic stagnation was beginning to manifest itself. Although the above mentioned indicators are associated with an increase in homicide risk, the problem is much more complex than that as is illustrated by Barnett, Kleitman and Larson (1975). The latter authors have shown that sizeable increases in risk could be detected in each of the nation's 50 largest urban places by the early 1970's. The black population is concentrated in a selected number of the nation's larger urban centers. Thus we will choose a sample of such places to focus our attention on as we attempt to arrive at a series of strategies designed to ameliorate the threat of homicide risk. The Status of Risk in Large Urban Environments To illustrate the heightened risk of homicide vicitimization in the nation's larger black communities, we selected the 14 largest ghetto centers for emphasis. Almost one-third of the nation's black population resided in these urban places in 1980 although some cities had begun to experience signs of population loss (Rose, 1982). Among the group of cities, all but St. Louis continued to maintain base populations in excess of 250,000 in 1980. The latter center had also satisfied that threshold of population size ten years earlier. Economic decline and the subsequent shift of economic activity out of the larger and older central cities have resulted in both a reduction in the attraction of selected centers to non-metropolitan migrants and large scale movement of central city black residents to inner suburban municipalities within the same metropolitan area. In some instances, selected segments of the black population remaining in the least attractive central cities or parts of central cities represent a residual population. Thus the 188 dynamics of population movement might be expected to have an impact on risk in centers where there is the most economic decline. Dimensions of Urbanism and Risk Both size of place (Mayhew and Levinger, 1976) and density (Gove, 1979) have been previously cited as having strong links to risk of victimization such that these urban dimensions are often employed as predictors of risk. Others conclude that urbanization itself is directly associated with risk (Parker and Smith, 1980). Yet there continues to be an absence of consensus on the role of urbanization of risk. How the various dimensions of urbanism operate within the context of a large territorial black population is unknown since most ecological-oriented researchers of homicide seldom investigate at that scale. Yet Bullock (1955) suggested more than a generation ago that the physical isolation of these communities combined with the lesser interest of the police in what occurred within the communities helped support lifestyles that were risk promoting. Support for the previous position is less than robust, however, since it was associated with a single case, Houston. More than 30 years have elapsed since the observations supporting that case were made. Therefore, we are basically operating in a vacuum insofar as it relates to a sound understanding of the environmental issues that contribute to heightened homicide risk in the nation's larger black communities. Risk Levels in Primary Target Communities By 1980, homicide risk in the black population had declined almost 10 points below the level of 44 per 100,000 registered ten years earlier. At the same time, blacks now accounted for a smaller percentage of total victims (48%). The trend toward a decline in black victimization dominance was first observed in 1974 and continued throughout the remainder of the decade. Yet the national trend has not manifested itself across all environmental types within the black community as a review of the cities within our sample reveals In'our 14 sample communities, risk was higher in 1980 than in 1970. A mean risk level of 55.8 per 100,000 was recorded for these cities, which is approximately 40% higher than the national rate for black homicide. Thus major urban centers represented more dangerous environments at the end of the period than at the beginning, and smaller urban environments and non-metropolitan environments afforded somewhat greater safety. The cities selected to illustrate the seriousness of risk represent all regions of the country, but there are more cities from the South and North Central regions. Among the cities in the sample, all but Atlanta had higher homicide risk in 1980 than in 1970. Risk among the black 189 population in the sample demons the mean (see Table 1). St. Lo at the upper end of the risk co and Baltimore are at the lower cities, the task of achieving t 1990 is clearly within reach. high side of the mean, however, heroic challenge. Yet, it is n by a substantial margin even wh the upper end of the continuum. trates wi uis and L ntinuum; end For he target For those th e task ot imposs en risk 1 de variation around os Angeles are found Memphis, Washington, the latter three level of risk by communities on the will constitute a ible to reduce risk evels are found near 190 Table 1 Total Risk and Black Homicide Risk in a 14-City Sample (per 100,000 persons in the population) City________ Total Risk Black Risk 1970 1980 1980 New York 14.5 23.9 46.8 Philadelphia 20.0 26.4 51.8 Baltimore 25.8 28.6 40.6 Washington 26.4 36.1 39.9 Atlanta(x) 51.3 38.6 46.3 Memphis 16.2 24.2 37.7 New Orleans 21.7 37.5 56.8 Dallas 29.1 33.6 55.3 Houston 25.7 37.1 45.9 Cleveland 34.5 45.7 76.8 Detroit 34.5 45.7 59.3 Chicago 24.8 28.6 45.1 St. Louis 39.9 48.8 91.6 Los Angeles 15.1 33.8 87.3 Source: Klebba, A. Joan, "Homicide Trends in the United States, 1900-1974," Public Health Reports, May-June 1975, Vol. 90, No. 3, p. 204. (x) Was not included in the above source. The 1980 rates were computed by the author based on information secured from the 1980 FBI Monthly Homicide Reports. 191 In 1970, homicide risk in Atlanta's black community was recorded at 82 per 100,000, but had declined to 46 per 100,000 by 1980. Thus a 44% decline in risk was registered over a decade. Just how this decline was engineered is not clear, but it was indeed achieved. Risk may have been reduced in Atlanta because of the development and introduction there of a domestic crisis intervention center in 1974. In fact, the risk of family victimization was lowered from approximately 21 per 100,000 to approximately 9 per 100,000 during the interval. In some ways, it appears that the Los Angeles risk trajectory experienced the inverse of that just described for Atlanta. In the nation's larger central cities, elevated risk is more serious than elsewhere. Furthermore, it is in those environments that risk will have to be lowered substantially if the suggested goal of risk reduction is to be achieved. Although the environments where elevated risk prevails are easily identified, the role played by these environments on risk is extremely complex. Most ecological assessments of risk usually employ some measure or combination of measures of economic well-being as predictors of risk (Blau and Blau, 1982; and Messner, 1983). Measures of social disorganization (Levy and Herzog, 1978) and, in some instances, measures of culture or surrogates of culture are employed (Gastil, 1971; and Lofton and Hill, 1974). But the strength of explanation of these measures varies from one set of investigations to another largely as a function of scale, i.e., regional, metropolitan, central city, etc. Although these investigations sometime produce conflicting results, they do provide a point of departure for considering the problem. Yet they are concerned, at least directly, with only a single dimension of the problem—the environment. On the other hand, individual level assessments of homicidal behavior often ignore the environmental context where lethal confrontations arise. Thus if we are to successfully pursue the proposed goal of lower risk, we must somehow integrate what we know about individual level behavior and the role of the environment in abetting those behaviors we wish to discourage. Furthermore, we must be able to assess the environmental contribution at various scales of resolution, e.g., macro, medial, and micro. The task of operationalizing the contribution of the various environmental scales on risk is not an easy task. But we prefer starting with the environment as a means of attempting to address those forces that impel selected individuals to engage in acts of lethality. Homicide Risk and the Macro-environment If we begin our discussion at the macro scale, we must focus our attention on that set of forces which has significantly altered the character and quality of American life over the past 30 to 35 years. During that period, economic growth has become less dependent on manufacturing, thus reducing the need for unskilled blue-collar workers. 192 Moreover, this economic growth has increasingly become dependent upon service activity, which has drawn a disproportionate share of women into the work force. Also, ghettos have expanded enormously in most of the nation s larger metropolitan areas; and violent crime rates have increased by more than 250% (Harris, 1981). These and other changes have led to value shifts and the emergence of alternative lifestyles. The Role of Changing Core Values Nevertheless, according to Spindler (1977), the basic core values have remained intact even though the degree of support for them may have been altered. Yet, Spindler concurs with Rokeach (1978) that value shifts are evident and that a set of emergent values has in some instances replaced traditional values. It appears that the speed at which these shifts in the strength of American core values have taken place is in part tied to the size and complexity of individual urban environments. Thus those places previously identified as representing the nation's primary black communities would probably represent places where the adoption of emergent values would occur early. But the speed of adoption is likely to be influenced by the share of the population that was socialized in a more traditional context, e.g, the non-metropolitan South. Therefore, the share of the population present in each urban place in 1960 that previously lived in a non-metropolitan setting should be expected to have an impact on values finding support within the context of the family setting. Furthermore, we would expect southern cities to show more resistance to the adoption of emergent values than would their larger non-southern counterparts where these values are more likely to originate. It is at the macroscale that the stage appeared to be set for promoting a new round of heightened individual risk of homicide victimization. The changing character of the economy lowered participation rates of blacks in the labor force especially those of black males who were thought to serve'as a catalyst in promoting value shifts. At the same time young black "baby boomers" were reaching maturity during a period of cultural instability, a situation that was more likely to promote the acceptance of emergent values When this is coupled with blacks having greater amounts of discretionary time because of declining job opportunities in the manufacturing sector, it is logical to expect an elevation of risk. One interpretation of this set of events was recently described in the following way: Manv millions of our citizens are trapped in inner-city ghettos 15 little hope of escape. They are being warehoused in wel- fare-supported prisons; they have become a pro^ively more isolated subculture that is totally alienated from the demands " a world dominated by computers, telecommunications, and more recently, robotics (Lesse, 1984). Although the above interpretation might be somewhat overdrawn, a recent interview conducted by Claude Brown 193 (1984) with a group of Harlem youths clearly brings home the impact of value shifts on the elevation of homicide risk at the microscale. Brown, who wrote the celebrated essay Manchild in the Promised Land more than 20 years ago, contends that "Manchild 1984 is the product of a society so rife with violence that killing a mugging or robbery victim is now fashionable" (Brown, 1984, p. 44). One young 16-year old told Brown: "You know, you take their stuff and you pop [shoot 'em" (Brown, 1984, p. 44). If that Harlem youth's attitude is widespread, then gratuitous risk can be expected to increase if we are unable to intervene. The previously cited changes in the nature of the economy and the subsequent corresponding changes in values and norms are critical in understanding both the phenomenon of risk elevation and the attempts to lower risk. What is apparent is that America in the post-industrial era is a qualitatively different place than that which prevailed prior to the occurrence of structural changes in the economy. The impact of the foregoing changes, which went unnoticed by most observers until the late 1950's, have now become apparent at the macro scale. Their micro scale response, however, is less well understood. This is especially true in terms of how these changes have had an impact on the nation's black communities. That lack of understanding partially results from the tendency of social science to focus on individuals and groups without seriously considering what role the environment plays. Because of this tendency, social commentators frequently express puzzlement when attempting to explain the levels of violence that characterize the behavior of individual subgroups in the population, e.g., blacks and selected other minorities. Risk and National Policy Another reason for beginning our assessment of risk at the macro level rather than at the micro level is related to our interest in policy formation. National policy is thought to have various impacts upon risk, many of which are only weakly understood. But it is thought that national policies directed at economic development, manpower development, educational support, illicit drug trade, and gun regulation may influence the level and prevalence of risk in ways that may not have been fully examined. Each of these broad policy areas, as well as others that have been omitted, should be explored to determine their influence on both vulnerable individuals and groups. This would indicate the extent to which they might be implicated in promoting the development of dangerous environments. 194 Risk and Economic Decline Under the circumstances, we would expect risk to be highest in places where economic decline is most advanced. It is in these environments that selected individuals are most vulnerable to the twin impact of declining economic opportunity and the adoption of values that weaken both family and community bonds. A review of the homicide rate prevailing in those cities upon which we have chosen to focus our attention, and which are also included in Nathan and Adams' (1976) assessment of central city hardship, provides some support for our position. In cities where the hardship rates were low, e.g., Los Angeles and Houston, homicide rates in the black community were lower than those in black communities where hardship rates were high, e.g., Newark, Cleveland, Detroit, Atlanta, and St. Louis. Bradbury, Downs and Small (1982) indicate that the central cities which had high rates of hardship continued to lose ground between 1970 and 1975. By mid-1970, each of the primary central cities that had a high rate of hardship was characterized by continuing decline and severe distress. At the latter date, however, Los Angeles—which ranked low on the hardship index five years earlier-was now included in the column of distressed places. Likewise, there was an increase of approximately 100% in homicide risk in that city's black community during the decade. Thus a limited review of central city hardship and economic decline lends some support to our position on the relationship between risk and economic decline. If this position holds it would appear that heightened risk—and hence the level of seriousness of homicide as a primary cause of death—would be essentially confined to large central cities undergoing economic decline. We do not choose, however, to oversimplify the relationship between economic decline and risk for in some instances high risk levels prevailed in sel^d J?^* communities where decline was not evident. The situation just described might be partially related to the implied homogeneity often imputed to the concept of homicide In fact homicide represents a multi-dimensional concept, the concurrence of which can be motivated by a wide range of stimuli. Megargee (1981) recently reviewed the battery of murder typologies that are currently in vogue. He was careful to point out that they range over a number of dimensions, including both internal and external motivations. The Changing Nature of Homicide A dichotomous categorization scheme, however, will allow us to address the basic motivations leading to most acts of lethality. The conflict/non-conflict or expressive/instrumental schema is sufficiently simple and enables one to more readily understand the circumstances that led to the fatal confrontations. Until recently, conflict-motivated homicide occurring as an outgrowth of 195 angry confrontation represented the unchallenged dominant pattern. But during the most recent decade, in selected environments, that dominance has been challenged by non-conflict-motivated patterns. Ye*-, the precise role that economic decline has played in altei ng the motivational mix is less than clear at this time. ved the values and the decline of traditional values. If this is : fact the case, we would expect the homicide mix to vary as function of the rate at which individual places undergo a value transformation. Also, we would expect the level of risk to partially parallel changes in the economic well-being of the most vulnerable segments of the population. Nonetheless, it is unknown if these two instigators of homicide behavior exhibit a differential sensitivity to similar levels of economic change. Our primary contention relative to the general obser\ increase in risk nationally is that economic decline in the central city promotes increased environmental stress. This is especially noticeable in environments where economic dislocation is most severe. How people adapt to stress varies among sub-populations based on age, level of educational attainment, marital status and social position. Therefore, a different set of responses should be expected on the basis of the previously identified status variables and the severity of neighborhood stress levels. But, more importantly, as it relates to the previously identified homicide-dichotomy, is the role of intervening variables (see Mooney and Brenner, 1984) that influence the adaptive process. One such intervening variable that has paralleled the rise in homicide levels in selected black communities is the increased risk of robbery victimization. It was previously noted that nationally the risk of robbery tripled between 1965 and 1975 (Cohen, Felson and Land, 1980). Cohen, Cantor and Kluegel (1981) attribute the observed change in risk of robbery victimization to changes in the pattern of routine activities on the part of the American population in response to structural changes in the economy. Thus the foregoing authors tend to support an opportunity theory of victimization. The observed increase in the level of robbery victimization during the late 1960's and early 1970's disproportionately involved black offenders and black victims. Evidence of the increase in black involvement in the robbery rate was not randomly distributed across urban places. Rather, it was initially concentrated in selected non-southern central cities. This involvement in robbery is viewed as evidence of the growing importance of acquisitive values among black youth, as well as their willingness to pursue these values through involvement in illicit activity. Thus greater 1% participation by blacks in activities leading to robbery suggests that black youths are more willing to engage in behavior on a level and frequency uncharacteristic of similarly-aged cohorts in the past. It is through the adoption of these values that the structure of homicide victimization was altered (Rose, 1984) and that non-conflict-motivated homicide increased in relative importance. Traditionally, homicides in black communities across the nation have been motivated by angry confrontations among two or more individuals. These confrontations grew out of interpersonal conflict, e.g., domestic conflict, drunken brawls, acts of revenge, etc. But during the most recent era (post-1965), traditional motivations have declined in relative importance such that stress levels may fail to satisfactorily explain differential homicide risk from one urban setting to another. Moreover, because values which support violent behavior are thought to be diffused across space and time (Fischer, 1980), it is quite possible that the differences presently observed between selected non-southern and southern central cities simply represent a lag effect. The intervening variable that we have introduced as possibly being an important contributor to the changing character of homicidal behvior, as well as to the changing level of homicidal risk, is thought to be culturally motivated. Value shifts, which promote exaggerated interest in individual accumulations of items of temporary merit and an orientation that leads to self-indulgence, are likely to have a negative impact on communities where economic trauma is most intense. Both Harris (1981) and Brown (1984) provide examples of this. In assessing the situation, the former writer points out: "While the AFDC mothers did not actively encourage their sons to enter the drug trade, everyone recognized that a successful drug dealer could become a very rich man" (Harris, 1981, p. 134). Brown, in describing the Harlem youth of today, says this: They appear driven by, or at least obsessed with, a desperate need for pocket money that they cannot possibly obtain legally. They possess an uncompromising need to be able to 'rock' (wear) a different pair of designer jeans at least twice a week, or even a different pair of ordinary pants twice a week (Brown, 1984, p. 38). Structural Shifts and Homicide Type During the past 20 years, the nation has undergone a major transformation especially as it relates to the nature and character of work and the range and diversity of lifestyles that have become available to a large share of the population. The quality of life of most Americans during that period showed signs of improvement. Although this was a time of substantial progress, there was also evidence of negative feedback occurring alongside positive change. The changing norms and values previously described had an impact on individual sub-groups in the population in different ways, largely dependent upon their position in the social order. 197 The element of negative feedback—viewed here as most central to our concern—is the growing loss of empathy among individuals in society. This loss of empathy frequently expresses itself in predatory behavior. Such behavior is exemplified by the notable increase in non-conflict-motivated homicide since 1965. Although this trend is evident in national aggregate homicide statistics, it is most visible in large urban centers with mature and growing ghettos. An increasing number of ghetto residents are at greater risk of being homicide victims in general, and being victims of non-conflict-motivated homicide in particular. If we are to effectively design policies that will significantly lower homicide risk in the nation's larger communities, we must be sensitive to changes that are occurring not only in terms of elevated risk but in changing homicide structure. Policies that might effectively lower risk associated with conflict motivation may not address the circumstances which promote an increase in the level of non-conflict motivated homicide or vice versa. But because these structural differences are not randomly distributed, it will be necessary to specify in what environmental settings one or the other variant is likely to be the more important promoter of increased risk. As early as 1970, there was evidence of the growing importance of non-conflict-motivated homicide to total risk. Smith and Parker (1980), who label this type of homicide as non-primary and inlcude within it gangland slayings insitutional homicide, felony homicide and suspected felony homicide, showed that at least one-third of the aggregate homicide rate in New York, New Jersey, Illinois, Michigan, Missouri, and California could be attributed to it. States that were less urban and/or had a much more traditional orientation were less likely to show signs of significant non-primary growth. Barnett, Kleitman and Larson (1975) employed magnification ratios to illustrate the rate of growth in homicide risk during the late 1960's. These ratios reveal that risk generally increased most rapidly in urban centers located in states where non-primary homicide was a major contributor to rising risk. Structural Change and Implications for Prevention: The Case of Polar Environments The impact of the differing environmental context on homicide structure and subsequently upon risk can be illustrated by employing two polar cases. To highlight the problem, we chose cases represented by the aggregate black community in New York City and Birmingham, Alabama. These communities are thought to best represent examples of environments that have been differentially influenced by the cultural changes described earlier. In the former community, change has advanced rapidly. Change in the latter community has occurred much more gradually. Many current anecdotal essays provide some support for 198 viewing New York's Harlem in particular as an environment where community disorganization has reached its apex. In Harlem, we would expect the traditional pattern of black homicide to be less evident than in the southern city of Birmingham. In fact, we would expect to find risk in New York's black communities more often to be felony based. We would, however, expect risk in Birmingham to be largely rooted in interpersonal conflict. Luckenbill (1984) describes these two polar orientations to violence as character coercion and instrumental coercion. He suggests, for instance, that a policy on handguns may have a positive effect in altering risk associated with one but not necessarily with the other. Basic Differences Between the Communities Since the size of the black population in these two communities varies enormously, we chose to examine the structure of victimization for a single month in New York and for a full year in Birmingham. The number of blacks who became homicide victims in November, 1980, in New York City (61) is quite similar to the number of blacks who were homicide victims (63) in Birmingham during the entire year of 1980. Information describing the structure of victimization prevailing in Birmingham's black community in the 1930's and early 1940's was previously described by Harlan (1950) and is employed as a base for measuring structural change. Harlan's study demonstrates that during the earlier era more than 90% of black homicides in Birmingham grew out of interpersonal conflict involving honor and self-respect. Thus, in the language of Levi (1983), Birmingham's pattern of homicide evolved out of conflict between "lovers" and "adversaries." Birmingham and the Traditional Pattern Based on the FBI's monthly homicide reports, the pattern in Birmingham appears to have changed very little. More than four-fifths of the homicides at the more recent date appear to have been motivated by conflict. One-third of the conflict-motivated homicides took place within the family; two-thirds involved friends and associates. Only one-sixth of victimizations in the southern city involved individuals for whom the relationship was unknown. Although the level of risk prevailing in Birmingham's black community in 1980 was high (39.8 per 100,000), it was on the low end of the risk continuum that characterizes risk in the nation's larger black communities. Therefore, the forces that stimulate traditional and non-traditional patterns of victimization are disproportionately skewed toward elevating risk along traditional lines. Thus Birmingham's black community has been very slow to adopt emergent values, resulting in the continued dominance of traditional violence-provoking stimuli. Or, the buffers that repel negative feedback in association with the adoption of emergent values have been unusually effective. 199 New York City and the Rise of the Non-traditional Pattern New York City, unlike Birmingham, Alabama, is a non-traditional environment where change takes place rapidly. Many forms of unconventional behavior originate in New York, and these are eventually diffused and adopted elsewhere in the nation. Here the nation's largest black community, Harlem, is located. It is also here that a much higher than average number of blacks annually become victims of homicide. Therefore, we expect the pattern of homicide victimizatipn, if not the pattern of risk, to differ substantially from that observed in Birmingham. Evidence suggests that the modal pattern currently observed in New York City was not evident until the late 1960's. This position finds support in both the qualitatively- and quantitatively-oriented literature. For instance, Claude Brown (1965) indicated that the tradition of fighting that he grew up with in Harlem during the 1940's and early 1950's was dead by 1957. It is that tradition which social scientists who are advocates of a subculture of violence often employ to support their position. In addressing the issue, Brown says the following: As I saw it in ay childhood, most of the cats I swung with were more afraid of not fighting than they were of fighting. This was how it was supposed to be because this was what we had come up under. The adults in the neighbor- hood practiced this. They lived by the concept that a man was supposed to fight (Brown, 1965). Brown further contends that the basic principles of the behavior of the black male revolved around his money, his woman, and his manhood; and interference in any of these areas could lead to a killing. According to Brown, however, these traditions were weakened by the growing presence of drugs. Thus the knife gave way to the gun just as spontaneous killings gave way to intentional killings. Employing a quantitative rather than a qualitative orientation, Swersey (1977) carefully examined police records describing the changing circumstances surrounding homicidal behavior in Harlem during 1968, 1973, and 1974. He found that assaults with a gun that were motivated by intent rather than by spontaneous disputes increased several fold over a five-year period. Moreover, he thought much of this increase was connected to drugs. Thus although Brown thought drugs weakened the proclivity of blacks to fight in defense of traditional values, Swersey views drugs as an intervening variable supporting the rise in number of assaults with a gun. The growth of Harlem's irregular ecnomy (Ianni, 1974)—involving drugs, prostitution, gambling, and petty thievery--no doubt did much to set the stage for a new round of activities that would lead to a change in both the pattern and risk of homicide. By 1970, the homicide rate in Harlem was 130 per 100,000; and the pattern showed increasing male dominance, a decline in intersex victimization, and a rise in lethal encounters not motivated by conflict. According to Helpern (1977), this pattern had reached epidemic proportion by 1972 when 57 homicides were 200 committed during a single week; and an unheard of 14 homicides were committed during a single day in New York City. The Connection Between Drugs and Guns Helpern largely attributed the previously dsecribed epidemic to weak drug laws. Yet Haberman and Baden (1976) demonstrated a strong association between abuse of narcotics and deaths caused by guns during this period. Guns represented the modal weapon employed by substance abusers. Alcoholics were found more likely to have died as a result of a stabbing. But the latter victims were also more likely to have been over 40 and to have grown up outside New York City. According to Halpern (1977), the most dangerous environments during this period were concentrated in a few selected locations: Harlem, East Harlem, and the South Bronx. Thus the changes observed in New York City during the late 1960Ts and early 1970's should be expected to lead to a pattern of victimization that could be easily distinguished from that observed in Birmingham's black community at the same time. Unknowns—an Emerging Category A review of the FBI's monthly data from 1980 that describes homicides in New York for a single month reveals the continual evolution of the previously described pattern. If we assume that none of the 27% of the victims whose race was unknown was not black, then blacks constituted 44% of the victims during the month. The single, most frequently occurring relationship between the victim and offender was an unkown one, and the circumstances surrounding the event were undetermined. Swersey (1977) identified events dsecribed in this manner as "apparent murder." We have classified these events as suspected felonies although Reidel (1984) suggests that some unknown deaths may turn out to be non-felonies on case closure. The declining clearance rate in those places where unknowns constitute a substantial number of the total events seems to support our intuitive decision to put such events in the category of instrumentally-motivated acts. These events represented apoximately 60% of total black victimizations in our sample month. Furthermore, when stranger victimizations are added to the mix, the total felonies and suspected felonies rise to more than 70% of the total victimizations. Thus interactions not motivated by conflict appear to represent the dominant interactions leading to homicide in New York City's black community. On the basis of these two polar examples, we have attempted to loosely illustrate the implicit impact of economic change and subsequent value shifts on the pattern of homicide victimization in two disparate black communities. If we are to realistically approach the problem of reducing the risk of homicide in the nation's larger black communities, then we should first consider the 201 role of the macro environment and its impact on the growing complexity of factors contributing to risk. To fail to do this would lead us to recommend a series of piecemeal solutions based on conditions at a single point in time rather than attempting to intervene in a temporal process. Prevention and the Micro Environme Studies of homicide seldom investigate th environment where most lethal interactions occ those investigators who employ an ecological a study of homicide seldom choose as the unit of neighborhood of victim's residence. Neverthel scholars who pursue the defensible space appro study of violence frequently do choose the nei scale as the appropriate setting for such inve (Newman, 1972; Taylor, Gottfredon and Brower, in attempting to articulate a set of preventiv designed to reduce risk at this level, we are a limited base of research. Fortunately, our own orientation (Rose an 1981) to assessing risk has been conducted at Moreover, at this scale, the results of lower easily observed. But the strategies necessary in risk at this spatial level are somewhat amb untried. Therefore, our rather brief treatmen microscale differences in risk are likely to b descriptive than prescriptive. Nevertheless, is the goal, and we will continually try to id workable strategies. nt e micro ur. Even pproach to the analysis the ess, some ach to the ghborhood stigations 1984). Still, e strategies left with only d McClain, this scale. risk can be to intervene iguous and/or t of e more prescription entify Level of Risk in Neighborhoods The neighborhood-scale patterns observed in Detroit, St. Louis, and Atlanta reveal that the intensity of risk in selected neighborhoods far exceeds the level of risk prevailing in the black community in general (Rose, 1981-82). Those neighborhoods where risk exceeded the level describing the national aggregate black risk in 1970 were identified as substantial-risk neighborhoods. Neighborhoods satisfying this criteria were further subdivided into zones of intermediate and high risk. High-risk zones (> 100 per 100,000) most often coincided with zones of high stress where economic dislocation had reached its apex. Intermediate-risk neighborhoods, however, were more often zones where risk levels exceeded what one would normally expect based on socioeconomic indicators prevalent in those areas. In both Detroit and St. Louis, intermediate-risk neighborhoods frequently exhibited higher than normal risk elevations resulting from superimposing rising non-conflict risk upon traditional conflict-motivated risk. In such environments, models devised to explain traditional patterns of behavior tend to falter as the apparent relationship between stress and aggregate risk tends to break down. When components of risk were disaggregated in St. Louis and Atlanta, model performance varied as individual independent 202 variables changed greatly in their explanatory power as a function of neighborhood context. For instance, vacancy rates and percent young adult black males were significant predictors of risk in those neighborhoods where the relationship between the victim and offender was unknown. On the other hand, rates of acquaintance victimizations were accurately predicted in neighborhoods where a higher than average percentage of persons were divorced or separated and a higher than average percentage of males were unemployed. Generally, however, we were better able to explain levels of risk in high-risk neighborhoods than in intermediate-risk neighborhoods. Patterns of Victimization in Neighborhoods and Preventive Strategies The above findings suggest some tentative approaches to lowering risk, but they are simply just that—tentative. What is particularly troublesome in this instance is the increasing contribution of unsolved cases in selected neighborhoods and our lack of precise understanding of the motivation and circumstances that brought victim and offender together. Victims frequently, however, live in some of the city's most physically-blighted neighborhoods. In other instances, higher than average vacancy rates occur in non-blighted zones where the dynamics of population turnover has led to excessive movement. In the latter situation, more people moved out than moved in. In such instances, the socioeconomic mix of the neighborhood might be altered in ways that promote a predator-prey situation. Thus our indicator of increased risk of felony murder may be producing a mixed signal. Until we more completely understand the relationship between micro environment and risk, robust preventive strategies will be slow to evolve. What this suggests, however, is the need to eliminate unoccupied structures where people can conduct illicit activity and/or where predatory assaulters can hide. It is unknown where these environments simply represent locations where the irregular economy is dominant or where a residual population—with only a weak commitment to traditional norms—is disproportionately concentrated. Under the circumstances, we must gain a sounder understanding of these relationships in order to prevent the previously suggested preventive strategy from backfiring. One of the shortcomings of relying too strongly on ecological indicators is the tendency to confuse these aggregate markers with individual markers. For instance, most victims of homicide are described as poor in terms of a general characterization—even when they are found to have resided in non-poor neighborhoods. Although it is true that the income mix of victims is changing somewhat, the projected image of the victim is that of a person with low income. Yet when we observe that victims in non-conflict-mot^vated confrontations frequently reside in lower, middle-income districts—especially if the 203 victims are young adults—it suggests they may live in their parental home, which is often a non-poor household. This simply represents an example of the confused signals sometimes given off by data collected and analyzed at different scales of resolution. The Prevalence of Substantial Risk When we attempt to develop preventive strategies based on various micro-environmental assessments, shortcomings abound. Still, much can also be gained from this investigational approach. For instance, not only are we able to identify the intensity of risk at the neighborhood scale, but we can also delineate the prevalence of risk. As long as risk is concentrated or is thought to be concentrated in a few selected neighborhoods, the perception of aggregate danger is minimal. Until recently, substantial homicide risk in the larger black communities tended to be confined to neighborhoods characterized by low status and high levels of disorganization. Bruce-Briggs (1976) rationalized that because of the persistence of this pattern of risk, there was little justification for promoting gun control laws designed to reduce risk. But the data show, especially in the nation's larger black communities, that substantial risk has spread beyond neighborhoods which were previously the most common settings for elevated risk. Therefore, the prevalence of neighborhoods that had substantial risk, as a proportion of all neighborhoods in the black community, has increased over time. Thus the rise of substantial risk in environments beyond those identified as the most distressed is essentially related to the increased importance of non-conflict-motivated homicide to total risk. As substantial risk becomes more prevalent, so does gratuitous violence accompanied by feelings of neighborhood insecurity in residential spaces that were previously perceived as secure. Factors Contributing to Prevalence The prevalence of substantial-risk neighborhoods is apparently sensitive to a range of factors. Variations in the change in the existence of prevalence were recently observed in the black communities of Atlanta, Detroit, and St. Louis. In both Atlanta and Detroit, substantial-risk neighborhoods as a proportion of all neighborhoods in the black community declined. The inverse of this pattern was observed in St. Louis. Almost two-thirds (64.9%) of all neighborhoods in St. Louis were substantial-risk neighborhoods in 1970, and had experienced only a modest increase (67.3%) by 1980. A sizeable decrease occurred in the proportion of substantial-risk neighborhoods in the former two cities, with Atlanta experiencing the most notable change. At this point, explanations for the observed changes are mere conjecture. But it seems safe to assume that both changes in overall risk and changes in the structure of i 204 victimization are likely to influence neighborhood prevalence. In Atlanta, where risk decreased rather dramatically during the decade, so did prevalence. The thinning out of population in that city's high-risk neighborhoods might have resulted in a lessening of risk. Yet very high risk could still be observed in neighborhoods where risk was traditionally high. Thus substantial risk was concentrated in only about 30% of the neighborhoods in the black community. At the same time, only slightly more than 30% of the neighborhoods in St. Louis escaped the substantial-risk designation. Apparently, a different set of forces was at work in these communities that influenced the spatial spread in the pattern of risk. The Problem of Misclassification of Neighborhoods A primary shortcoming of this mode of analysis, and one that might encourage misleading interpretations at the neighborhood scale, is a reliance upon small numbers. A minimum of three homicide events is generally adequate to qualify a neighborhood as one of substantial risk. Therefore, random variations can result in a neighborhood being classified as one of substantial risk during any single year. Only if there is evidence of persistent risk can we be certain that labelling the neighborhood as "dangerous" is appropriate. This problem is further compounded by whether such labels are employed to identify the neighborhoods where the event takes place or the neighborhood where the victim or offender resides. We have chosen to follow the convention employed by local health departments, which is to assign all victimizations to the neighborhood of victim's residence. Although that practice has obvious shortcomings, it is generally thought to favor the kind of ecological analysis undertaken by social scientists. For those whose approach to the problem is more oriented to the physical environment, the location of event is likely to take precedence over either the location of the victim's or offender's residence. One advantage in employing victim's residence—beyond being able to compile information on several critical variables thought to indicate level of neighborhood distress—is the ability to record the movement pattern of the victimization vis-a-vis the site of the fatal confrontation. During an earlier period, it was safe to conclude that the site of fatal confrontation and the place of residence of the victim and offender were likely to have occurred within a close radius (Bullock, 1955; Pokorny, 196 ) Today, however, that radius is likely to have increased in response to both an alteration in the spatial dimensions of social networks and changes in the structure of victimization. # One might raise the question of how persistent linkages between neighborhoods of social origin and those of current residence have an impact upon spatial patterns of risk. Likewise, to what extent do offenders journey to physically 205 and/or socially remote neighborhoods in search of victims? Either of the above examples could lead to an increase in the prevalence of risk in locations that might be considered atypical substantial-risk neighborhoods. Persons engaged in assessing homicidal risk frequently overlook questions of the above nature. Moreover, the Brantigans point out, for instance, that the position Wolfgang assumed in his classic homicide study was "that the spatial pattern of offenses was of no importance, and was of local interest only to police" (Brantigan and Brantigan, 1981, p. 27). This position, however, was taken prior to the rise in relative importance of street victimization. Partitioning the Environment In today's context, growing attention is accorded the role of both the internal and external environment on risk of victimization. The former milieu is generally defined in terms of selective characteristics of internal residential space, i.e., density, crowding, housing amenities, and composition of household. Although a small battery of scholars has devoted much effort in attempting to specify the association between density and/or crowding and crime in general, the results are thought to be inconclusive (Harries, 1980). Yet, Grove, Hughes and Galle (1979) produced results that link crowding—at least—to a number of pathological consequences. This suggests these measures should not be summarily dismissed. Thus it is possible that this line of research may yet yield insight which bears upon homicide risk within physical, environmental settings. As street crime has increased in relative importance, so have attempts to manipulate the external environment in ways designed to both deter offenders and protect individuals from unwarranted physical assaults. The targets of much of this suggested physical manipulation are designs of buildings, patterns of street lighting, uses of arterial land, and boundary markers for neighborhoods. The impetus for much of the foregoing work is based on Newman's (1972) elaboration of the defensible space concept. Newman (1980) contends that the combination of low-income families and high-rise public housing projects provides just the kind of anonymous environment where the crime rate—especially the robbery rate—might be expected to increase. Newman attributes the increased risk of crime to the lesser possibility for surveillance, the number of entrances per unit, the height of the building, and the lack of a clear demarcation of public and private space. Environments possessing these characteristics have become less attractive to people with low income as is attested by the high vacancy rates in some of the nation's largest high-rise public housing projects. One city, St. Louis, demolished a housing project that was thought to manifest all of the ills generally attributed to such environments (Rainwater, 1973). Yet we are unaware 206 what effect the elimination of this structure has had on homicide risk per se in St. Louis' black community. The large increase in robbery that occurred during the previous 15 years has indirectly led to an increase in homicide risk. Therefore, it is important that we are able to modify environments which favor this activity. Block (1977) indicates that 60% of all robberies in Chicago occur on the street, at night, and during the work week. At least 2% of armed robberies, especially those committed with a gun, lead to death (Block, 1981). So although the significance of defensible space approaches is thought to be modest at this point (Taylor, Gottfredson and Brower, 1984), their true impact on reducing homicide risk at the neighborhood scale is as yet uncalculated. Individual Attributes and Risk of Victimization The discussion of context, both macro and micro, has dominated our thinking in the development of this essay. The primary reason for this is related to the very limited emphasis placed on the contribution of the environment generally. In no way, however, are we suggesting that the role of the individual is not central to our understanding of the factors which contribute to heightened levels of victimization. But, according to Megargee (1981), it is extremely difficult to predict individual behavior in a vacuum. Likewise, most known attempts to lower risk are individually centered and fail to consider the role of the environment on behavior (Rappaport and Holden, 1981). Therefore, there is a serious need to view the individual within a specific context if we are to increase our understanding of the factors that contribute to risk. The perspective of the individual that follows will cite only a limited number of attributes which are thought to heighten the prospects of risk enhancement in the nation's black communities. The factors selected for a tentative review are as follows: a) gender; b) age; c) schooling; and d) lifestyle. Although these represent individual level attributes, they are also independent. And, as an aggregate set of characteristics, they are directly influenced by the ever-changing complexities of the environment. Sexual Differences and Risk The risk differential that characterizes homicide victimization, at least among blacks, is strongly based on gender. Therefore, any program designed to reduce the level of black victimization should, of necessity, address the issue of gender-induced risk levels. The ratio between black male and black female victimization ranged between 4 and 5:1. Even though there is a large discrepancy between the risk levels of males and females, black females have higher levels of risk than those of whites of both sexes. In 1975, the risk of victimization for black females was 14.7 per 100,000; and risk for white females was less than 3.0 per 100,000 (Shin, 1981). Thus blacks are more 207 vulnerable to homicide risk regardless of gender. Although it is quite clear that black males are at much greater risk of victimization than are any other race-sex category, by 1975 the level of risk among the latter group exceeded 70 per 100,000 (Shin, 1981). Since there was an estimated 15% decline in aggregate black risk between 1975 and 1980, one would assume that risk for black males should have declined to a level approaching 60 per 100,000, whereas risk for black females should now be less than 13.0 per 100,000. There is no evidence, however, to indicate that the observed decline was uniformly distributed between the sexes. On the contrary, it appears that risk for black females declined more rapidly that risk for males. Moreover, there is growing evidence that a convergence is occurring in the general mortality experience of black and white women, at least in selected age groups (Ueshima, 1984). But it dos not appear that this carries over to the area of homicidal risk There is evidence, however, which suggests that risk for black females is declining in response to a series of evolving group-specific attributes. Among these, one obvious change is a slowdown in group-specific marriage rates. No doubt this has had the effect of reducing the risk of spousal homicides. The limited direct evidence we have regarding the foregoing pattern is that recently presented by Zimring, Mukherjee and Van Winkle (1983). Those authors demonstrate that there was a 100% reduction in intersexual killings in Chicago between 1965 and 1980. The phenomena they described was dominated by black participation. In Chicago, the situation during the most recent 15-year period differs sharply from that describing Detroit in an earlier period (Boudouris, 1971). Familial homicide was described as a very substantial contributor to the total pattern of homicide in Detroit during the two generations preceding 1968. It is unfortunate that it required a slowdown in family formation to have a positive impact on intersexual and/or spousal homicide risk--if this is in fact the case. A more precise examination, however, of the impact of declining marital rates is needed in order to confirm its association with decreasing levels of intersexual victimization. Continued sex role differences lead to sizeable differences in risk of mortality in general (Waldron, 1983) and risk associated with violence in particular (Day, 1984). It is said that males take more risks and engage in more daring activities than do females, a pattern which enhances the general risk of death. One specific pattern that stands out, in terms of its likely impact on homicidal risk, is the reported greater use of illegal psychoactive substances by males (Waldron, 1983). A recent review of the sexual composition of drug-related deaths in one precinct in New York City partially sustains the latter point as well as sustains the 208 more general notion of sex role differences. The data describing the gender characteristics of the victims in the New York precinct revealed that 90% were male (Heffernan, Martin and Romano, 198Z) . It appears that primarily males engage in drug trafficking and that the product market has a male bias as well. Therefore, in black communities where drug trafficking has gained a major foothold, one should expect a further alteration in the sex ratio of victimization. Age and Vulnerability to Risk Another ascribed characteristic, age, shows a definite association with changing levels of victimization. Moreover, rates of victimization are generally shown to be highly sensitive to the population's age structure within local areas. Some have attributed the general upward movement of crime rates in the United States to the coming of age of the "baby boom" population. That group assumes a decline in general incidence will occur as the "baby boom" cohort reaches maturity. Wolfgang (1978) has indicated that the 1980's will have lower crime rates in response to the relative decline in the proportion of persons 15-24 years of age, whom he indicates are the most criminogenic. Thus it is generally thought that people born between 1945 and 1959 are primarily responsible for the sharp escalation in homicide risk that occurred between 1965 and 1974. But, how much can we expect homicide risk to be reduced in the nation's black communities as this age group matures? The Magnitude of Changes in Demographics on Risk A number of analysts of levels of homicide victimization have attempted to specify how much changes in levels of risk were directly attributable to changes in the population's demographic structure. Barnett, Kleitman and Larson (1974) have deduced that less than 10% of the increase observed between the mid-1960's and early 1970's could be attributed to changes in demographics. Turner, Fenn and Cole (1981), however, estimate that 25% of the increase in homicides during the previous 15 years could be attributed to an increase in population growth. Thus there are differences in opinion regarding just how important demographic factors have been in promoting the observed changes. Like Wolfgang, the latter authors anticipate a sizeable downturn in criminal victimization as an outgrowth of the decline in the number of persons in the crime-prone years. Moreover, by the late 1970's, there was evidence that rates of homicide victimization were declining among young black males. The Prime Target of Vulnerability As previously indicated, the 15-24 year old male population has ben singled out as the group at greatest risk, both as victims and as offenders. Block (1975) Illustrated this point in his study of Chicago that covered 209 a nine-year period. He showed that 15-24 year old males accounted for 25.1% of the increase in victims and 35.6% of the increase in offenders. Nevertheless, by 1978, risk within this population group had declined from 102 per 100,000 in 1970 to 78 per 100,000 in 1978 (Morbidity and Mortality Weekly Report, 1983). The Center for Disease Control has shown, however, that regional and size of place differences in risk within this population continue to persist. The North Central region, which includes a number of mature ghetto centers embedded in traditional manufacturing-center economies, maintained levels of risk that were almost twice those in the nation's other three regions. Moreover, the observed pattern indicates widespread differences in level of vulnerability of high-risk age groups as a function of the environmental context. Several writers have attempted to explain why young adults in general and blacks in particular become involved in interactions that tend to catalyze risk. It has been said that this population is more frequently the target of aversive events (Turner, Fenn and Cole, 1981), and therefore is more often inclined to resort to aggression as a means of terminating the event. The quick resort to aggression is thought to be related to an increase in consumption of alcohol (Waldron and Eyer, 1975; and Weiss, 1976); drug abuse and addiction (Weiss, 1976); impulsive rage and a breakdown in respect for societal institutions (Waldron and Eyer, 1975). The severity of the impact of the above changes on young blacks recently led Gibbs (1984) to describe them as an endangered species. The changes ushered in during the 1960's are thought to have led to an elevation in risk across all age groups in the black population. But the 15-24 year old group has been identified as the most vulnerable to aversive life events that underlie noxious behavior. Yet when we examine data from individual places, it is obvious that although risk is generally attenuated among the population previously identified, other age groups have not been unaffected. Individual places reflect the structure of victimization, and the position of an individual place on the structural continuum is likely to greatly affect age differences in risk. Risk in those communities where traditional patterns dominated, until recently, showed the highest levels of risk were concentrated in the 25-34 year old populations rather than in the 15-24 year old groups. Likewise, the decline in risk with age was much more gradual in southern communities than in non-southern communities. Emerging patterns of risk, however, have led to an intensification of risk at younger ages. Zimring (1984) recently inferred that a further intensification of risk among adolescent populations is limited only by the younger segment of that group's lack of access to handguns. Thus an altered orientation of young adults, coupled with more frequent encounters with aversive events and increased access to handguns, has made this population the most 210 vulnerable to homicide risk. Differences in Age Structure of Victimization Among Urban Places Limited evidence based on observations from a small number of urban centers illustrates the variations in age structure noted above. Where risk decreased sharply during the prior decade, its impact was most readily observedin the under-35 age group, with principal targets of vulnerability represented by persons between 20 and 24 or 25 and 29. Thus, although risk among younger black populations was said to be declining, ths was not the case in St. Louis and Los Angeles. Yet there was substantial decline in risk among this age group in Atlanta and Houston. The obvious distinction between places where growth or decline was evident was the rate of change in non-conflict-motivated homicides. For instance, in Los Angeles, 74% of the homicide victims who were black males aged 20 to 29 in 1980 appear to have been involved in a felony or suspected felony. In Atlanta, only slightly more than one-fifth of the same age victims appear to have been involved in non-conflict encounters. Thus it seems, on the basis of a limited number of observations, that the growth or decline in risk for young adult black males is strongly associated with their involvement in criminal behavior either as a victim or as an offender. Prevention Must Be Targeted to Assist All Age Groups If the goal of reducing aggregate risk to a specific level is to be attained, preventive efforts will need to be sensitive to the relationship between age and risk of victimization. Although specific age groups are more vulnerable than others, preventive efforts should not ignore the plight of those populations who, on the surface, appear less vulnerable. Therefore, although elderly people and young teens have less risk of victimization, they are made more vulnerable with each increase in the aggregate risk. Nevertheless, the growing concentration of risk among young adults poses some very difficult problems of prevention involving a broad array of public institutions at the local level and the need to consider various policy options at the national level. Schooling and the Identification of High-Risk Individuals .8 there is a literature which clearly shows risk increases as level of educational attainment decreases. But from the perspective of prevention, it appears that schooling as a social process is more likely to shed light on the identification of individuals for whom risk is likely to be levated. So although level of educational attainment as a ummary measure may logically predict risk in an ecological e is 211 context, schooling is believed to represent a more valuable approach to understanding individual risk. Unfortunately, only a very limited aspect of this complex subject will be addressed here. The following question might be logically raised: why consider schooling in an effort concerned with lowering the risk of homicide victimizaton? One might answer that schools have continuous contact with the individual for a longer period of time than does any other institution outside the family; that schools provide an environment where values and norms are adopted; and that schools also help launch the individual on a path that will greatly influence his or her life. But beyond the previous global effects, in the school various behaviors can be observed that provide insight in assessing potential for risk. Thus, on the basis of observed and recorded behavior in the school setting, early warning signals can be detected. According to Robins (1968), truancy is the most critical ealry warning signal that is a predictor of homicide among young black men. Likewise, behavior in school provides evidence of a longitudinal pattern of psychosocial development that could be useful in predicting risk. Academic Performance and Risk First, one might examine the academic performance of those individuals who at an early age persistenly demonstrate signs of scholastic marginality. Such signs are usually strong indicators of future alienation from schooling in particular and eventually society in general. The observed indicators might derive from various sources, some of which represent deficiencies in nurturance and others in nature. Lawrence (1975) also points out the importance of noxious or stressful experiences encountered by individual children that lead to developmental problems. Any of these deficits can lead to what Brummit (1978) has chosen to identify as dyssocial children. The point is that these children are more likely to grow up to be individuals who have a higher potential risk than do their peers whose scholastic performance falls within the normal range. To be sure, one must carefully search out the source of difficulty that is the basis for the observed academic performance. Once this has been identified, corrective measures should be instituted. It is well-known that the previously cited students formally come to the attention of school authorities early. School systems have developed an elaborate labelling system to identify the kinds of deficits these pupils are experiencing, e.g., emotional disorder or behavioral disorder. Yet to what extent are these problems perceived to impair the functioning of individuals outside the school environment? And to what extent are efforts made to ameliorate the problems so as to help these people function both in and out of school. Too often we read—after the fact--that some young 212 person was involved in a wanton act of violence in which one or more individuals have been fatally victimized. At that point, it usually surfaces that the offender and/or victim was known to be suffering from some mental disease, a condition that had no doubt persisted for some time. It is also likely that signs of such disorders were previously identified in school. But it is unlikely that the student received anymore than minimal help. Nevertheless, it is possible that had there been early intervention--given previous signals—the violent event may never have occurred. Other Dimensions of Schooling and Risk As a measure of a sstudent's successful involvement in scholastic activity, academic performance is only one dimension of the schooling process, albeit the most important. Other dimensions also have an impact on the student's perception of the value of participating in schooling as a social activity. Extramural programs play an increasingly important role in school. These programs influence the student's perception of school as well as influence the probability of risk. Many programs provide selected students with an opportunity to showcase their athletic skills and/or performing talents in ways that lead them to feel good about themselves. This often occurs even when their academic performance leaves much to be desired. Thus not only does schooling provide pupils with a set of long-term advantages, it also offers short-term boosts in self-esteem. From our perspective, such programs often keep students in school who might otherwise withdraw and take their chances in the job market. As indicated earlier, students who are poorly prepared and who drop out of school before graduation have a higher risk of being involved in fatal situations. Students who are unable to find a niche for themselves in the school's social structure also are likely to have a weaker commitment to scholastic values. Moreover, students who have poor academic records and who also have no athletic or performing talent are likely to find school less attractive than a range of alternatives. Even though students may continue to go to school, their attendance is based almost totally upon attractions that have little to do with schooling per se. Yet there comes a time when even those attractions are no longer adequate to prevent the students from dropping out of school. If Robins is correct, the greater rate of withdrawal prior to graduation leads to higher levels of anticipated risk. At present, high school drop-out rates average about 20% (Elwood and Wise, 1983). But in many inner-city enviroments, the rate of withdrawal is somewhat higher than in other school districts. In these inner city settings, the incongruence between the value of the school and those of the individual exacerbate the potential for risk. 213 Scholastic Behavior of a Sample of Victims Some of the previous statements can be validated by reviewing the scholastic records of a sample of homicide victims in a large, midwestern city almost a decade ago. Although the number of observations under review is small (N=25) , these observations strongly parallel those made earlier by Robins (1968). All of the victims were born in the state of residence and were under 30 at the time of death. The sample (22 males and 3 females) is congruent with the expected sexual ratio in environments where non-conflict homicide is beginning to show incipient dominance or near dominance. Moreover, the single largest number of victims (25%) matriculated at one high school prior to graduation. The most striking feature describing our sample was the high rate of poor scholastic performance and/or anti-social behavior. Of the 25 students in the sample, three-fourths were suspended or withdrew before graduation. Most dropped out of school before reaching the eleventh grade, and some already had minor scrapes with the law. Many students also participated in special educational programs designed for pupils with specific deficiencies. Yet only one student in A closer review of the individual records revealed that members of the group were frequently absent. As a group, these students possessed several attributes previously identified as promoting elevated risk. Within the school setting, if markers or cues can be identified that indicate a student is at high risk of victimization, then an all-out attempt should be made to develop programs that might help lower individual risk. Lifestyles and Risk A final attribute that should be given some attention in any formal effort aimed at lowering risk is the set of lifestyle orientations adopted by people who live in substantial-risk environments. Lifestyle, as a concept, has been variously defined; and it appears that some ambiguity is still associated with its use. Some writers tend to focus on behavioral propensities that are an outgrowth of status differences (Silverstein and Krate, 1975), and others tend to emphasize world view and a subsequent set of activity patterns (M. Gottfredson, 1981; and Smith 1982). Still others tend to associate the concept with one's personal orientation to life and the corresponding behaviors which project that image to one's associates (Mancini, 1980; Sobel, 1981). Needless to say, although each of these approaches is somewhat different, each involves a routine set of activities. These activities are generally conducted within a highly circumscribed sociocultural milieu where the ultimate objective is to define oneself for others and to elicit responses that are satisfying to the ego. Lifestyle, as we have chosen to define it, simply represents the way individuals choose to spend their leisure and the rewards they anticipate achieving as a result of 214 participating in certain activities. Prior socialization experiences, personality, value orientation, and social status are highly intertwined. Moreover, they do much to abet the adoption of a specific lifestyle. Several lifestyles adopted by black residents of substantial-risk neighborhoods often prove troublesome and are important in the elevation of risk. Among the more troublesome are those dsecribed as hustlers (petty), drug-culture oriented, swingers, and crime-oriented. These are not mutually exclusive styles, but they do tend to represent the primary orientation of individuals. The goals of people who have these lifestyles range from simple survival, e.g., petty hustler, to enjoyment of the good life, e.g., local drug dealer. A careful review of the circumstances associated with patterns of homicide victimization illustrates the pervasiveness of the previously identified lifestyles on risk. During the most recent period of escalation of risk, illicit drugs had the greatest effect on the growth of troublesome lifestyles. There is a substantial literature that details the role of alcohol upon victimization and the extent to which the victim and offender were found to have been drinking prior to the homicide. Scholarly support for the drug-homicide nexus, however, is not as evident. We do know that drug addicts tend to be younger than alcoholics (Haberman and Baden, 1978). Thus we should anticipate that persons involved in lifestyles where drug use is an important behavioral trait tend to be young (<35); to be involved in a social network made up of similar others (Flaherty, Kotranski and Fox, 1984); and to be willing to engage in a wide variety of unconventional behaviors to satisfy their need for drugs (Ben-Yehuda and Nachman, 1984; and Strug and others, 1983). Although the literature on the connection between drugs and victimization and their association with risk is yet sparse, a few works have surfaced that provide some modest insight into the relationship. For example, a paper based on evidence of the presence of drugs in the bodies of a sample of Detroit victims in 1973 is most revealing (Monforte and Spitz, 1975). That study found that 43% of the victims under 35 had used drugs. Likewise, the authors distinguished deaths related to drugs, i.e., drug trafficking, from the broader definition of user deaths that might have involved alternative motivations. In environments where the lifestyles of young black adults are similar to those observed in Detroit, elevation of risk can be expected. ....... . . .. If we are to be able to substantially lower homicide risk, we must develop the know-how to discourage black youths from adopting self-destructive and narcissistic lifestyles. These lifestyles do much to expand the range or radius of risk. As we suggested in a prior section of this essay, embryonic evidence of a given lifestyle orientation often'manifests itself in both social and academic behavior 215 at school. Therefore, baseline cues are available and indicate the need to initiate efforts to direct the individuals who have been identified toward more socially acceptable lifestyles. Such persons should be directed to legitimate avenues in order to assist them in achieving more socially rewarding lifestyle. It is crucial, however, that we be able to determine if the primary support for destructive lifestyle orientations originates in the home; in the micro social environment outside the home, e.g., the neighborhood; or if it basically emanates from contact with the national cultural media, e.g., television. It will be much more difficult to minimize the influence of the latter two elements; but with parental support, it is possible to reduce their effect. Without parental support during crucial stages in the developmental process, the battle might well be lost. Instead of describing individual lifestyles in depth, we have simply chosen to broadly identify a select few. Those chosen for review appear to contribute substantially to both individual risk as well as to neighborhood risk. The most troublesome lifestyles frequently are connected to illicit drugs. Therefore, unless we are able to have an impact on the flow of drugs into the community, drug-related deaths are likely to increasingly contribute to risk (Heffernan, Martin and Romano, 1982). This has led to the "hit man" becoming a central figure in drug related deaths in cities where drugs are a major growth industry (see Dietz, 1983). But the availability of drugs is essentially a function of perceived demand. Thus the curtailment of the adoption and practice of the previously described lifestyles is directly related to our success in discouraging drug use and the effectiveness of treatment for current users seeking a cure. SUMMARY In this essay, we have attempted to draw attention to the population that is at the greatest risk of dying as a result of homicide. We are aware that the target population has long maintained higher than average levels of homicide risk. But it is only since the middle 1960's that elevated risk has occurred as an outgrowth of non-traditional elements. From a historical perspective, risk in the nation's black communities has been most often associated with intensely emotional confrontations growing out of a loss of self-esteem. According to some researchers, these confrontations—or at least the method selected to resolve them—were partially related to a minority-status effect. Others attributed the behavior to membership in an oppressed group. Still others attribute the early pattern of homicide involving blacks, which originated in the rural South, to the impulsiveness of blacks (Heilbrun, Heilbrun and Heilbrun, 1978). Data now confirm that the traditional contributors to risk are being challenged by emerging situations which simultaneously heighten risk, at least in selected environments, and makes prevention of homicide more 216 difficult. It has been suggested that a specific goal be established and that every effort be made to lower risk to the target level. We have illustrated, however, that risk varies greatly from one major urban area to another. Therefore, individual places may not attain the national aggregate goal over the same temporal span. Nevertheless, based upon our previous description and discussion, it is assumed that rational goals for individual places, along with the appropriate target interval, can be established. Thus our discussion simply attempts to lend insight into the nature of the problem, but falls short of providing a specific blueprint for prevention of homicide. The complexity of the problem is such that no single set of recommended strategies is likely to prove successful across the broad spectrum of individual circumstances and environmental settings known to contribute to risk. For that reason, we suggest that homicide be treated like the nation's other major killers, that is to become the target of ongoing basic research which is adequately funded. Most of what we currently know about homicide is not the product of research aimed at prevention. Rather, it is basically an outgrowth of testing heuristic structures related to disciplinary and/or intellectual development. Current need, however, dictates that there be greater integration between theory and practice. We have only touched upon a selected set of issues that have an impact upon homicide risk in the nation's larger black communities. There are some highly important issues that we decided not to pursue in this paper, largely because of the lack of unanimity regarding their role in risk. If the initial effort becomes bogged down in the merit or lack of merit of individual techniques or policy choices, a stalemate could emerge in getting any prevention program off the ground. Issues that readily come to mind and that tend to divide otherwise collegial colleagues are as follows: 1) What constitutes an appropriate policy on guns? 2) What are appropriate guidelines for sentencing? On the issue of guns, some of the nation's leading authorities on the upsurge of risk believe that the easy availability of guns should be considered the prime explanatory variable (Block, 1975; Farley, 1980; and Zimring, 1984). Zimring recently expressed the intensity of his feelings on the issue by stating as follows: "Even though blacks are disproportionately victimized by homicide, Jesse Jackson has not gotten up and yelled, 'Get the goddam guns out of your house.' That's public health education" (Meredith, 1984). Surely, not all agree with that position. Needless to say, however, the problem of guns must be resolved if risk is to be brought unier control. The second thorny issue that tCxids to divide scholars with a professional interest in -he topic is the practice of sentencing, especially that involving the idea of selective incapacitation. The latter practice is being urged by some criminologists who have found that the most serious crimes 217 are committed by a very small percentage of a given age cohort. They reason that if this small band of criminals were given more severe sentences, taking criminal history into account, risk of victimization could be substantially lowered. Others, however, view this move away from the desert model as troublesome, largely because of the high rate of false positives associated with attempts to predict future criminal behavior (von Hirsch, 1984). Neither of these issues is likely to disappear, but one can be sure that they will continue to be discussed and suggested as appropriate techniques. From an intuitive perception, it appears that those who support a strong regulatory policy for guns are opposed to some of the evolving sentencing guidelines, especially preventive incapacitation. But it also appears that supporters of the latter option tend to be less enthusiastic about attempts to regulate availability of guns. We have attempted to provide a way of looking at the problem that will assist in designing effective prevention efforts. Our approach takes into consideration both external macro- and micro-risk stimuli. Further, it suggests that if these are ignored, lowering levels of risk will be more difficult. Also critical to our understanding of risk is the role played by individual attributes. We acknowledge the shortcomings of an essay of this type, but contend that it represents one valid approach toward a better understanding of the problem. From this holistic perspective, serious thought can be given to decreasing the risk of victimization in the nation's larger urban communities. If a rational set of prevention strategies is not actively promoted, an increasing proportion of the nation's black population will be relegated to a permanent status of having unacceptably high levels of risk. They will also be subject to the subsequent emotional turmoil of knowing that danger is ever present. 218 References Cited Barnett, Arnold, Kleitman, Daniel 3. and Larson, Richard C. On Urban Homicide: A Statistical Analysis," 3ournaJ__of_C£J_mj_naJ_ 3ustJ_ce, vol. 3, 1975, pp. 85-110. Ben-Yehuda, Nachman, "A Clinical Sociology Approach to Treatment of Deviants: The Case of Drug Addicts", Dr_ug_an d_AJ_c o h o i_Depe n d e n ce , vol. 13, 1984, pp. 267-282. Blau, 3udith R. and Blau, Peter M., "The Cost of Inequality: Metropolitan Structure and Violent Crime", American_Socj_oj_ogj_caJ_ Review, vol. 47, Feb. 1982, pp. 114-129. Block, Richard, VJ. o J_e n t _C£ j_me , Lexington Books, Lexington, Mass., 1977. Block, Richard, "Homicide in Chicago: A Nine-Year Study (1965-1973," Ill£_2°yiIl^I_£l_^limill^I_L^_in.d_Qlimil22i£Sy.- vo1* 66» No* ^' 1976' pp. 496-510. Block, Richard and Zimring, Franklin E., "Homicide in Chicago, 1965-70", 3 o urn a J__o .f _Re s e a r c h_iH_C r j_rne _a n d _De iip^u e n cy, vol. 10, No. 1, 3anuary 1973, pp. 1-12. Bradbury, Katharine L., Downs, Anthony and Small, Kenneth, Urban D££lil!£_^n£LIbe_Future_o^ The Brookings Institution, Washington, D.C, 1982. Brantingham, Paul 3. and Brantingham, Patricia L., Environmental QHim!H2i2£Y-' SaSe Publications, Beverly Hills, Cal., 1981. Brenner, M. Harvey and Mooney, Anne, "Unemployment and Health in the Context of Economic Change", Socj_aJ__Scj_ence_and_Med|cj_nex vol. 17, no. 16, pp. 1125-1138 Brown, Claude, Manchj_J_d_in_the_Promj. sed_Land, Signet Books, New York, 1965 Brown, Claude, ManchHd_jm_HarJ_em, Sept. 1984, pp. 36-44 and 54 and 76-77. Bruce-Br iggs, B. , "The Great American Gun War", The_PubJ_ _j_c_J_n teres t, no. 45, Fall 1976, pp. 37-62. Brurrmit, Houston, "Socialization of Dysocial Children", The American 3 o u r n a J__o f _P s xc h o a n a j_x s i s , vol. 38, pp. 31-40. Boudouris, 3ames, "Homicide and the Family", 3ournaJ__of _Mar r j_age and_the_Famn_y_, Nov. 1971, pp. 667-676. Bullock, Henry A., "Urban Homicide in Theory and Fact", 3ournaJ__of Criminal Law_a n d_Cr hmn o J_ogx ■ vo1- ^5> 1955- PP- 565-576. 219 Center for Disease Control, "Violent Deaths Among Persons 15-24 Years of Age-United States, 1970-1978", Mor b^d£ty_and_Mor ta]J_ty_ Wee k ix.Repo r t , vol. 32, no. 35, Sept. 9, 1983, pp. 453-457. Cohen, Lawrence E., Felson, Marcus and Land, Kenneth C. , "Property Crime Rates in the United States: A Macrodynamic Analysis", ^m£li£an._l°u.IHii_°l_§2£i°i2SX* vol. 86> no. 1, 1980, pp. 90-118. Cohen, Lawrence E., Cantor, David and Kluegel, 3ames R., "Robbery Victimization in the United States: An Analysis of a Non-Random Event", Socj_aJ__Scj_ence_Quar_terJ_y_, vol. 62, no. 4, Dec. 1981, pp. 644-657. Day, Lincoln H., "Death From Non-War Violence: An International Comparison", Socj_aj__Scj_ence_and_Medj_c^ne, vol. 19, no. 9, 1984, pp. 917-927. Dietz, Mary L., K|IIing_for_Prof|t, Nelson Hall, Chicago, 1983. Ellwood, David T. and Wise, David A., "Youth Employment in the 1970's: The Changing Circumstances of Young Adults", Arnej-jcan Famiii2i_ ai?^_il}2_5£2I!2!ry» eds., Nelson, Richard R. and Skidmore, Felicity, National Academy Press, Washington, D.C. 1983, pp. 59-108. Farley, Reynolds, "Homicide Trends in the United States", Demography* vol. 17, no. 2, May 1980, pp. 177-188. Fischer, Claude S., "The Spread of Violent Crime from City to Countryside, 1955 to 1975", RuraJ__Soc\_oJ_og£, vol. 45, no. 3, 1980, pp. 416-434. Flaherty, Eugenie W., Kotranski, Lynne and Fox, Elaine, "Frequency of Heroin Use and Drug User's Life Style", ^e£J_can_3ou£na_j__of^ QlHS_iH^_Al£o!}2l_^H i £, vol. 10, no. 2, 1984, pp. 285-314. Gastil, Raymond D., "Homicide and a Regional Culture of Violence", Arn£li£an._§2£i£l°iai££l_?:2Xiew, vol. 36, 3une 1971, pp. 412-427. Gibbs, 3ewelle T., "Black Adolescents and Youth: An Endangered Species", ^e£ican_3ou£naJL_o^_0£^ho£^chj_a^t£y, vol. 54, no. 1, 3an. 1984, pp. 6-21. Gottfredson, Michael R., "On the Etiology of Criminal Victimization", I!2e_3ou£naJ__o^_C£Jmj_naJ__Law^ vol. 72, no. 2, 1981, pp. 714-726. Gove, Walter R., Hughes, Michael and Galle, Omer R. "Overcrowding in the Home: An Empirical Investigation of its Possible Pathologi- cal Consequences", ^e£J_can_Socj_oJ_ogj_caJ__Rev_i_ew, vol. 44, February 1979, pp. 59-80. Harlan, Howard, "Five Hundred Homicides", 2^IIIai_oi_ClimiJD^J_Iraw an!LClimi.D°i°iLy> vo1* ^°* 1950- PP- 736-752. Haberman, Paul W. and Baden, Michael M. , Alcohol Other Drugs and Violent De_at h_. Ox ford University Press, New" Yo7T<"T97in 220 Harries, Keith D., C£J_me_and_t^he_Eny^£onmenjtf Charles Thomas Publisher, Springfield, Illinois, 1980. Ha r r i s , Ma r v i n, Ame£j_ca_New^_The_An_t h£0 p oi£gy_o^_a_Changj_ng CuJ_Uj£e, Simon and Schuster, New York, 1981. Heffernan, Ronald, Martin, 3ohn M. and Romano, Anne T., "Homicides Related to Drug Trafficking", Fede£aJ__P£obat^on, 1982, pp. 3-8. Heilbrun, Alfred B. 3r., Heilbrun, Lynn C. and Heilbrun, Kim L., "Impulsive and Premidated Homicide: An Analysis of Subsequent Parole Risk of the Murderer," The_3ou£naj__of^_C£J_mJ_naJ__Law_and 9limiH2l2SI• vo1- 69> no- *• 1978, pp. 108-114. Helpern, Milton (with Bernard Knight), Aujtopsy, St. Martin's Press, New York, 1977. Holinger, Paul C. and Klemen, Elaine H.,"Violent Deaths in the United States, 1900-1975", S oc j_a J__S C£e n ce_and_Medj_c j_ne , vol. 16, 1982, pp. 1929-1938. Ianni, Francis A.3., BJ_ack_Ma f^j_a, Simon and Schuster, New York, 1974. Klebba, A. 3oan, "Homicide Trends in the United States, 1900-74", PHkii£_y£a!i!l_Re.EOIii- vo1- 90* no* 3* May-3une 1975, pp. 195-204. Lesse, Stanley, "Slavery - Pos t industr ial Style", Ame£J_can_3ou£naj_ o_f_Psy_chothe£apy, vol. 38, no. 1, 3an. 1984, pp. 1-3. Levi, Ken, "Homicide As Conflict Resolution", Devj_ant_Behavj_o£, X2ii_i2.-11*9.> PP- 281-307. Levy, Leo and Herzog, Allen, "Effects of Crowding on Health and Social Adaptations in the City of Chicago", U£ban_Eco^ogy, vol. 3, 1978, pp. 327-354. Lofton, Colin and Hill, Robert H. "Regional Subculture and Homi- cide: An Examination of the Gasti1-Hackney Thesis", Ame££can §2£i2i2Si£ii_E2Xi£w, vol. 39, Oct. 1974, pp. 714-724. Luckenbill, David F., "Character Coercion, Instrumental Coercion and Gun Cont ro 1" , The_3ou£na_[_of _Ap£J_j_ed_Behavj_o£aJ__Sc_i_ence , vol. 20, no. 2, 1984, pp. 181-192. Mancini, 3anet K. , St£ategJ_c_Stxl£s, Coping in the Inner City, University Press of New England, Hanover, N.H., 1980. Mayhew, B. and Levinger, R., "Size and the Density of Interaction in Human Aggregates", ^e££can_3ou£naJ__of^_Soc_i_oJ_ogy, vol. 82, no. 1, 1976, pp. 86-110. Megargee, Edwin I., "Psychological Determinants and Correlates", C£im£naJ__V£oJ_ence, eds., Wolfgang, Marvin and Weiner, Neil A., Sage Publications, Beverly Hills, California, 1982. Meredith, Nikki, "The Murder Epidemic", Sc J_en ce_84, vo 1 . 5, no. 10, Dec. 1984, pp. 42-48. 221 Messner, Steven F., "Poverty, Inequality and the Urban Homicide Rate", Crmuno\_ogy_, vol. 20, no. 1, May 1982, pp. 103-114. Monforte, 3. R. and Spitz, W. U., "Narcotic Abuse Among Homicide Victims in Detroit", 22Hin§I_2i_E2I2n-i!£_5£ie-n.£e-» voi* 20» Jan# 1975> pp. 188-190. Nathan, Richard P. and Adams, Charles, "Understanding Central City Hardship", PoJ_^t j_caJ__Sc£ence_Quarte£j_x> vol. 91, no. 1, Spring, 1976, pp. 47-62. Newman, Oscar, Conrrnu n j_ t y__of_J_n t e£e s t, Anchor Pres s/Doub leday , Garden City, New York, 1980. Newman, Oscar, Defensible Space, Macmlllan Co., 1972. Pokorny, Alex D., "Comparison of Homicide in Two Cities," Journal of Criminal Law, Criminology and Police Science, Vol 56, 1965, pp. 479-487.--- --------;----------------------- Parker, R. and Smith, M. , "Deterrence, Poverty and Type of Homicide", ^2lican_3ou£naJ__o£_SocJ_oJ_ogy, vol. 85, 1979, pp. 621-629. Rappaport, 3ulian and Holtfen, Karen, "Prevention of Violence: The Case for a Non-Specific Social Policy", Vj_oJ_ence_and_the_V£oj_ent l_nd LviduaJ , eds., Hays. 3. Ray, Roberts, Thomas K. & Solway, Kenneth S. SP Medical and Scientific Books, New York, 1981. Riedel, Marc, "Blacks and Homicide", Ih£_C£irrj_i_naJ__3us tj_ce_Sys t^em an d_B j_ac k s^, ed., Georges-Abey ie , Daniel, Clark Boardman Co., New York, 1984, pp. 51-60. Robins, Lee N., "Negro Homicide Victims - Who Will They Be?", Ila.n^.Za£li22» June, 1968, pp. 15-19. Rokeach, Milton, "Change and Stability in American Value Systems, 1968-1971", PubJ_j_c_Opj_nj_on_Qua£te£j_x, Surrmer 1974, pp. Rose, Harold M., "The Changing Spatial Dimension of Black Homicide in Selected American Cities", 22u.Ln.aI_2i_§!2Xil.2nm2Iliai_5Xli2mi- vol. 11, no. 1, 1981-82, pp. 57-80. Rose, Harold M. , "Black-On-Black Homicides: Overview and Recom- mendations", The_Crmu naJ__3ust j_ce_Syslem_and_Bj_acks , Clark Boardman Co., New York, 1984, pp. 61-74. Rose, Harold M. and McClain, Paula D. , BJ_ack_Homj_cj_de_and_t^he yikaH_§HXil2Hm£n_!• NIHM Final report, 1981. Shin, Yongsock, Dj_f_f^£entj_aj_s^__i_n_H2mi£±d£_in_l!2£_yJQiied_Sjtajtes^, J_930-J_97 5£__A_Demog£a2hj_c_Sjtudy, University Microfilms Interna- tional, Ann Arbor, Michigan, 1981. Silverstein, Barry and Krate, Ronald, Chj_J_d£e n_o^_^he_Da£k_Gh e t t^o , Praeger Publishers, New York, 1975. Smith, Susan 3., "Victimization in the Inner City", B££t£sh_3ou£naJ of_CrmunoJ_ogy, vol., no. 2, Oct. 1982, pp. 386-402. 222 Smith, M. Dwayne and Parker, Robert N., "Type of Homicide and Variation in Regional Rates", Socj_aJ__Forc.es , vol. 59, no. 1, 1980, pp. 136-147. Sobel, Michael E., L__es__y_e_and_Soc_____________• Academic Press, New York, 1981 . Spindler, George D.,"Change and Continuity in American Core Values", We_t he_PeopJ_e__Ame£_can_Cha_ac t e__and_Socj_aJ__Change " , ed . , D i Ren zo, Gordon 3., Greenwood Press, Westport, Conn., 1977. Strug, D. and Others, Hustling to Survive: The Role of Drugs, Alcohol and Crime in the Life of Street Hustlers", Unpublished paper, 1983. Swersey, Arthur 3., "Homicide in Harlem and New York City", Unpublished paper, Sept. 1981. Taylor, Ralph B., Gottfredson, Stephen D., and Brower, Sidney, "Black Crime and Fear: Defensible Space, Local Social Ties, and Territorial Functioning", 3ou£naJ__of_Resea£ch_£n_C££me_and 5e.iin2H2Il£X• vo1- 21> no- *» Sept. 1984, pp. 303-331. Turner, Charles W., Fenn, Michael R. and Cole, Allen M., "A Social Psychological Analysis of Vio_en__Behav_o£", Violent Behavior, ed., Stuart, Richard B., Brunner/Maze1 Publishers, New York, 1981. Ueshima, Hirotsugu, and Others, "Age Specific Mortality Trends in the U.S.A. from 1960 to 1980: Divergent Age-Sex-Color Patterns , 22Hin^I_2i_Chl2Qi£_5iA2ai£- vol. 37, no. 6, 1984, pp. 425-4 39. von Hirsch, Andrew, "The Ethics of Selective Incapacitation: Observations on the Contemporary Debate", C£_me_and_De__n_uenc_, vol. 30, no. 2, April 1984, pp. 175-184. Waldron, Ingrid, "Sex Differences in Illness Incidence, Prognosis and Mortality: Issues and Evidence", Soc_a__Sc_ence_and___d_c_ne, vol. 17, no. 16, 1983, pp. 1107-1123. Waldron, Ingrid and Eyer, 3oseph, "Socioeconomic Causes of the Recent Rise in Death Rates for 1 5-24-y r-ol ds" , Soc_a__Sc_ence_and Medj_C£ne, vol. 9, no. 7., 1975, pp. 383-396. Weiss Noel S., "Recent Trends in Violent Deaths Among Young Adults in the United States", Ame£lcan_3ou£na!_of_E2£derruo!ogy_, vol. 103, no. 4, 1976, pp. 416-422. Wolfgang, Marvin, "Real and Perceived Changes of Crime and Punishment", Daedalus, vol. 107, no. 1, Winter 1978, pp. 143-158. Zimring, Franklin E., "Youth Homicide in New York: A Preliminary Analysis", The_3 o u£nai_o f _LegaJ__S t u d j_e s , vol. 13, 3an. 1984, pp. 81-99 . Zimring, Franklin E., Mukherjee, S. K. and Van Winkle, Barrik, "Intimate Violence: A Study of Intersex Homicide in Chicago", The Unive£S£tx_of_Ch£cago_Law_Revj_ew, vol. 50, 1983, pp. 910-930. 223 Interdisciplinary Interventions Applicable to Prevention of Interpersonal Violence and Homicide in Black Youth Deborah Bouding Prothrow-Stith, M.D. Assistant Professor of Medicine Boston University/Boston City Hospital Boston, Massachusetts Interdisciplinary Interventions Applicable to Prevention of Interpersonal Violence and Homicide in Black Youth Interpersonal violence and its most devastating outcome, homicide, are endemic in urban black areas with low socio- economic indicators. Those who are most affected are young and male. The homicide rates for black men are seven to twelve times those for the general population. Homicide is the leading cause of death for black men ages 15-24 years at a rate of 72.5/100,000 (1) and for black men ages 25-44 years at a rate of 125.2/100,000 (2). Homicide rates for the general population have remained between 9-10/100,000 for the same time period (3). Non-fatal interpersonal violence does not result in the literal loss of lives, yet it occurs at rates that are at least a magnitude higher than homicide and likely represents an even greater overall cost to society. There is less ade- quate data on non-fatal interpersonal violence as it is more difficult to measure than fatal episodes and there is no legal requisite reporting of episodes. Emergency room and school data are the best source of rates for non-fatal inter- personal violence. However, these rates are still underesti- mations of the true rates because many episodes of interper- sonal violence are neither treated in emergency rooms nor do they occur in schools. The Northeastern Ohio Trauma Study calculated the inci- dence of cause-specific trauma in an area of population 2.2 million by collecting emergency room data for the year 1977. The study reported an assault rate of of 862 per 100,000 population. The overrepresentation of urban blacks of lower socioeconomic status was demonstrated in this study as well. The incidence rate for assault in the urban minority neigh- borhood was over twice the total incidence rate and up to six times the lowest neighborhood rate (4). The Statewide Childhood Injury Prevention Program, SCIPP, of Massachusetts, compiled data from the emergency rooms of fourteen communities statewide (Boston was not in- cluded and many of the communities were suburban). Older adolescents had higher assault rates than the Ohio study. The assault rate for 12-15 year olds was 73.6 per 10,000, the assault rate for 16-17 year olds was 164.8 per 10,000 and the rate for 18-19 year olds was 180.2 per 10,000 (5). School based data is as compelling. During the 1969-70 school year, Seattle Public Schools had 3.6 assaultive inju- ries per 1000 students (6). In the United States there are approximately 75,000 assaultive injuries to teachers a year at a rate of 35 per 1,000 (7). A November, 1983 publication from the Boston Commission for Safe Schools (8), reported a survey of four public high schools revealing that fifty 227 percent of the teachers and thirty eight percent of the students reported being victims of a school based crime during the school year. The overrepresentation of urban black students was evident in this report as well. Black students are suspended at rate of seventeen per one hundred compared to a rate of eight per hundred for white students according to this Boston report. A large number of the suspensions are for interpersonal violence. Weapon carrying behavior was also reported in this Bos- ton survey. Seventeen percent of the girls and thirty seven percent of the boys reported bringing a weapon to school at some point during the school year. Another Boston survey of predominately black tenth grade high school students revealed that forty percent of the students reported that they could get a gun if they felt they needed one (9). Homicide is a measurable form of interpersonal violence. Non-fatal violence is less reliably measured, yet occurs at significantly higher rates than homicide. The overrepresen- tation of young blacks is common to both. The severity and urgency of the problem for urban black communities dictates the need for appropriate and effective prevention strate- gies. Several factors have been shown to be associated with high rates of interpersonal violence. Low socioeconomic status was shown to account for the overrepresentation of blacks among homicide victims in a recent Atlanta study. Data that was corrected for socioeconomic status using the number of people per square foot of housing, no longer showed a racial bias (10)• In this study lower socioeconomic status was the factor associated with death by homicide. Urban black adolescents, overrepresented among the poor with unem- ployment rates of 40-60% (11) , are overrepresented among the victims of homicide. Other factors which teach or facilitate interpersonal violence are associated high rates. These include handguns, alcohol, television violence, early exposure to violence as a witness or victim and the rare instances of psychiatric or medical pathology. Developing effective prevention strategies requires concentrating on those factors which appear most amenable to change. Society has been unable to rectify the wrong of poverty or its race bias for blacks. Yet, teaching conflict resolution, reducing the availability of handguns, reducing the television violence viewed by children, or teaching non-violent disciplinary techniques to parents are examples of prevention efforts aimed at such factors. Reducing the rate of interpersonal violence and of black male homicide victims by 1990 as mandated by the Department of Health and Human Services1 priority objective (12) will re- quire interdisciplinary, multi-institutional strategies aimed at these factors. 228 The possibilities for such prevention strategies were greatly enhanced by the recent conceptualization of interper- sonal violence as a public health problem. Traditionally violence was viewed as only a law enforcement problem which limited both the professional expertise and the variety of institutions involved. This conceptual change, marked by the 1980 establishment of the Centers for Disease Control's Vio- lence Epidemiology Branch, not only enhances strategic possi- bilities but also is more appropriate given the "intimacy" characteristics of homicide. Though interracial violence, violence between strangers and violence that occurs during the commission of another crime predominate in the media, the predominate characteris- tics of homicide are the opposites of these. Eighty percent of the victims are the same race as their assailant. Fifty- eight percent of the victims of homicide knew their assail- ant, and twenty percent of the victims and their assailants are members of the same family. Fourty-seven percent of the homicides are precipitated by an argument, while only fifteen percent are related to the commission of a felony (13) . These "intimacy" characteristics of homicides resulted in the 1980 FBI Uniform Crime Report declaring that, "Murder is a socie- tal problem over which law enforcement has little or no control." A comparison between law enforcement data and emergency room data on assault rates furthers the case for a public health approach to interpersonal violence. Emergency room injury incidence rates for assaults were 3.8 times higher than those reported to the law enforcement officers for the same geographical area and time period (14). The Public Health Community is preparing to accept the inherent challenge with Surgeon General C. Everett Koop at the forefront. Dr. Koop has stated that, "Violence is every bit a public health issue for me and my succesors in this century as smallpox, tuberculosis and syphilis were for my predecessors in the last two centuries (15)." The traditional public health model of disease attributes the occurence of disease to complicated interactions between the environment, the pathogen (the agent that is responsible for the disease), and the host (the individual with the disease). Traditionally the public health model has been applied to unintentional injuries. The application of the model to violence prevention offers a particular challenge because of the intentional nature of violence related injuries. The public health interventions applied to other problems that have been most successful are those that have manipulated the environment and have had little dependency on changes in human behavior. Yet when applied to intentional injury environmental manipulations can be expected to be less effective. For example, a safety lock on the trigger of a handgun could be expected to prevent handgun related acci- dents, yet this intervention could not be expected to have that same effect on intentional shootings. Altering the 229 environment so as to prevent the occurence of a disease is often the most effective and cost efficient method of disease prevention and control. This is because manipulations of the environment require litle to no change in human behavior to be effective. The usage of child proof safety caps on medi- cations, the addition of flouride to the water supply and the use of safety devices in the design of many occupational tools are examples of such environmental manipulations. When applied to interpersonal violence and homicide, several envi- ronmental changes have been recommended. Tenants have out- lined recommendations for buildings with better lighting, without secluded areas and with increased security presence (16). Teachers and school administrators have recommended changes in the school structure and environment in order to reduce violence and delinquency (17,18). Eliminating handguns from the environment is a contro- versial homicide prevention strategy. Handguns are used weapons in sevety-five percent of the homicides. A causal relatioship between the availability of handguns and homicide has not been established. Guns are more lethal than knives or fists and potentially turn what would otherwise be a fight into a homicide. Eliminating handguns from the environment could decrease the number of homicides, yet the total numbers of episodes of interpersonal violence would not be expected to change. Altering the host (victim) and the pathogen (assailant) to prevent interpersonal violence and homicide is dependent on changing human behavior, which is more difficult than altering the environment. The goal in such manipulations is to make the host more resistant to the disease, as in vacci- nations and prophylaxis and the pathogen less virulent. Pre- venting interpersonal violence in urban young black men re- quires an appreciation of the distinct similarities between the victim and the assailant. Ruth Dennis, PhD in her work entitled "Profile: Black Male at Risk to Low Life Expectancy" compares three groups of black men ages 18-34 (19) . One hundred and sixty seven incarcerated homicide perpetrators, one hundred and thiry victims of serious assault (knife and gun wounds) and two hundred five randomly selected non-institutionalized black men obtained through household sampling. Social and psycho- logical profiles of each study participant were done and the three groups were compared. The victim and perpetrator groups were similar and dis- tinguishable for the control group, in that they had less education, had experienced more juvenile detentions, were more likely to carry a gun and were more likely to have been in jail before. In addition to having similar characteris- tics distinct from the control group, these two groups had more subjects to interchange the roles (victim vs. perpetra- tor) during the study. Because of this role interchange and 230 the similarities between the classically defined victim and prepetrator, when the public health model is applied to interpersonal violence the host and the pathogen are equal. Prevention strategies designed to make the individual less likely to involve in fighting are applicable to both. The legal distinction between victim and perpetrator remains valid yet the distinction is less valid when applied to prevention strategies. Several concepts which are critical to the development of prevention strategies result from the characteristics of interpersonal violence including, the intentional nature of the injuries: 1) The traditional public health model of disease is less applicable to interpersonal violence, yet it does supply a framework inwhich to discuss prevention efforts; 2) prevention strategies are dependent on the manipulation of human behavior to a larger extent than traditional public health injury prevention strategies; 3) environmental manipulations are expected to be less effec- tive when applied to intentional injury; 4) because handguns are more lethal than knives or fists, their elimination stands out as a potentially effective envi- ronmental manipulation assuming that the intent during a fight is to hurt and not to kill; and, 5) because of the potential lessened effectiveness of envi- ronmental manipulations, those prevention strategies designed to impact upon the individual by increasing his threshold for fighting are in the forefront and apparently are applica- ble to both the potential victim and perpetrator. Prevention strategies designed to increase an indivi- dual's threshold for violence are predominately education and behavior modification techniques. Teaching conflict resolu- tion and using role play to practice alternatives to violence are such strategies. Currently implemented school based programs which have been moderately successful with urban black adolescents are programmatic examples which are discus- sed further. The tenuous application of the host, pathogen and envi- ronment disease model is not the most significant gain from the conceptualization of interpersonal violence as a public health problem. Perhaps the most significant gain is the potential application of a multi-insitutional and interdisci- plinary model which has been applied to other public health initiatives. The national campaign to reduce smoking is an example of such an initiative. The media, health care insti- tutions, public schools, job sites, health fairs and county fairs became the source of educational information and mcen- 231 tive groups. Product labeling and advertisement restrictions were a part of the effort. This composite approach has led to a reduction in the number of smokers in this country. Much of this campaign model is applicable to interpersonal violence prevention when it is understood as a public health problem. THE BLACK ADOLESCENT Designing violence prevention strategies that are effec- tive with urban black adolescents of lower socioeconomic sta- tus requires an understanding of adolescence, of issues of race and of poverty. The discussion herein will not exhaust the theories of adolescent development or the impact of race or poverty on such. A general outline of these issues as applicable to the development of violent behavior is presen- ted. Adolescence is that period of dynamic physical and psy- chosocial maturaion which is the transition from chilhood to adulthood. The physical changes are the growth and develop- ment of puberty. The psychosocial changes include both cog- nitive maturation from concrete to abstract thinking and the mastering of specific developmental tasks. The major devel- opmental tasks are: 1) Individuation from family with the development of same sex and opposite sex relationships outside of the family (this results in the formation of a self identity separate from the family which necessitates experimentation with rela- tionships and experiences outside of the family and necessi- tates a transient period of narcisism or self love); 2) Adjustment to the physical changes of puberty with the development of a healthy sexual identity (often adolescents experience a transient stage of identification with sexual extremes, that is boys will indulge in a macho image and girls in images of extreme femininity); 3) Development of a moral character and a personal value system (which requires experimentation with different value systems; and, 4) Preparation for future work and responsibility. Failure to accomplish these tasks can result in signifi- cant dysfunction for the adolescent which can impair him as an adult. The tasks are accomplished in tandem and are the major requisites for healthy adulthood. The experience of poverty and of racism can significantly hinder the accom- plishment of these essential tasks. The development of a healthy self identity requires a sense of self-esteem and a healthy racial identity, both of which can be undermined by poverty and racism. Preparing for future work and responsi- bility is a meaningless enterprise when unemployment rates 232 are astonishingly high. Developing a sense of moral charac- ter and a functional personal value system is a least diffi- cult when television is one of the main sources of values. Viewing television violence is a significant contributor to the development of violent behavior patterns in children (20,21) . One of the most difficult problems facing service provi- ders for adolescents is that of defining normal behavior. Normal behavior for adolescents includes a variety experimen- tal behaviors which at other developmental stages would be abnormal. Defining normal is even more difficult in cases where there is a subcultural experience. Claude Brown in his literary work Manchild in _M Promised Land, describes such an experience: "...Throughout my childhood in Harlem, nothing was more strongly impressed upon me than the fact that you had to fight and that you should fight. Everybody would accept it if a person was scared to fight, but not if he was so scared that he didn't fight." The example clearly illustrates the dilemma. How much fighting is too much? ' When is it problematic? Many would agree that violence in self-defense is appropriate, yet if a homicide results would running not have been a better respon- se. On the other hand, in a violent world is it not healthier to defend oneself rather than be beaten or harrassed. Black adolescents suffer disproportionately high rates of both fatal and non-fatal violence. There are several required characteristics of adolescence which which make an adolescent more prone to violence. One such characteristic is narcissism. The narcissism is required for the adolescent to make a transition from family to the outside world. This required narcissism is responsible for the extreme self- conscious feelings of adolescents which makes them extremely vunerable to embarrassment. The adolescent feels that he is always in the limelight and on center stage. He is particu- larly sensitive to verbal attack, and it is nearly impossible for him to minimize or ignore embarrassing phenomena. Another adolescent characteristic that predisposes to violence is the transient stage of extreme sexual identity, or macho. Estab- lishing a healthy sexual identity requires transient stages of extreme femininity for girls and macho for boys. Macho is often synonymous with violent. The image of a coward is a deadly one for a male adolescent in this stage. Peer Pressure has been labeled the single most important determinant of adolescent behavior (22). This vunerability to peer pressure is a normal part of adolescence which faci- litates the accomplishment of several of the developmental tasks, yet it is a characteristic of adolescence which enhan- ces the predisposition for violence. If fighting is the expectations of peers as illustrated in Claude Brown's quote, 233 then an adolescent is least able to disregard those expecta- tions. Erikson (23) describes a societal moratorium from re- sponsibility that is necessary during adolescence to allow the requisite experiemental behavior to occur without compro- mise of future options. Thus, the adolescent is able to experiment with a variety of roles withou making a commit- ment. There is debate as to whether this moratorium occurs at all, yet many agree that in the situation of poverty it does not. The poor adolescent struggles with the develop- mental tasks without the protection of a moratorium. The black adolescent has to develop healthy racial iden- tity, in addition to the listed developmental tasks. Contact with racism results in anger that appears to contribute to the overrepresentation of black youths in interpersonal vio- lence. Psychologist Ramsey Lewis used "free floating anger" to describe anger not generated by a specific individual or event but from global factors such as racism, limited employ- ment options (24). This anger is the excess baggage that an individual brings to an encounter that lowers his threshold for directed anger and violence. This concept is helpful in that it attempts to account for the environmental and socio- economic factors and not label the individual as deficient. The anger is normal and appropriate. Violence prevention is therefore designed to change the response to anger to a healthier one, not to eliminate the anger. Factors associated with interpersonal violence include unemployment, poverty, low education achievement, exposure to television violence, alcohol consumption, weapon and handgun carrying are often factors which are more prevalent in poor urban black settings. The overrepresentation of black youth in the violence statistics appears to result from several factors which are characteristic of poverty. Violent behavior is learned behavior. The development and reinforcement of aggression and violent behavior in children is the subject intense study (25). Those factors which teach and reinforce violent behavior are more prevalent in urban, poor, black neighborhoods. Violence prevention programs which are appropriate for adolescents developmentally and which have a realistic cultu- ral context can be expected to be effective. Developmentally appropriate programs utilize peers in education and counsel- ing and reflect and understanding stages adolescents. The cultural context has to acknowledge the violence, racism and classism that many such adolescents experience. The problem of interpersonal violence among poor black adolescents has been long appreciated by frontline service providers and despite inadequate demographic data and an incomplete understanding of the causal factors, school and community based programs have been developed with moderate successes. 234 THE BOSTON YOUTH PROGRAM VIOLENCE PREVENTION CURRICULUM (26) The Boston Youth Program is a comprehensive health care initiative for adolescents funded by the Robert Wood Johnson Foundation. The health care services are hospital or clinic based and the health education prevention services are school based. A violence prevention curriculum developed for tenth grade health students is one of the health education services. The Boston Youth Program curriculum on anger and violence has been instituted in four Boston high schools and one community agency setting. To date approximately five hundred minority students have recieved the curriculum. The curriculum is designed to: 1) provide statistical information on adolescent violence and homicide; 2) present anger as a normal, potentially constructive emo- tion; 3) create a need in the students for alternatives to fight- ing by discussing the potential gains and losses from fight- ing; 4) have students analyze the precursors to a fight and practice avoiding fights using role play and video-tape; and, 5) create a classroom ethos which is non-violent and values violence prevention behavior. The prevention curriculum is specifically aimed at rais- ing the individual threshold for violence, by creating a nonviolent ethos within the classroom and by extending his repetoire of responses to anger. It acknowledges the exist- ence of societal and institutional violence and the existence of institutional racism. Students are not taught to become passive agents, but they are expected to claim anger and become intentional and creative about the responses to it. Anger is presented as a normal, essential and poten- tially constructive emotion. Creative alternatives to fight- ing are stressed. The classroom discussion during one ses- sion focuses on the good and bad results of fighting. The list of bad is invariably longer than the good, thus the need for alternatives. This exercise emphasizes that whether to fight or not is a choice and that the potential consequences are important to consider when making the choice. Role playing a fight is a unique part of the curriculum. During this session the students are asked to create a usual fight situation. The situation is analyzed for the build up or escalation phase, the role of the principal characters and the role of the friends on the crowd. Videotaping the role play is useful for discussions. Provocative behavior is 235 labeled and alternative behavior is discussed. The focus of the discussions is the demonstration and reinforcement of preventive behavior. The ten session curriculum has been evaluated using pre and post testing in one of the high school settings (27). This controlled study involved four tenth grade health class- es containing 106 students (approximately one third of the tenth grade enrollment for the school). Two classes (n=54) were assigned to the experimental group while the other two classes (n=52). The violence prevention curriculum was pre- sented to the experimental group, while the control students continued with the regular health curriculum. Both groups were evaluated by the same pre and post instrument approxi- mately ten weeks apart. The instrument tested for both know- ledge and attitudes about anger, violence and homicide. The experimental group had significantly higher post- test scores than the control group. There was no difference between the pre-test scores for the two groups. Knowledge scores accounted for more of the change than did the attitude scores. Though the change in attitude was significant with P <.01. These differences in scores represent the effect of the Violence Prevention Curriculum. Student questionaires were used to evaluate the curricu- lum. Eighty-seven percent of the students enjoyed or very much enjoyed the unit. Seventy-three percent of the students found it helpful with handling depression and sixty-three percent found it helpful in handling anger. This demonstration project shows that students receptive and enthusiastic about a curriculum on anger/homicide, and that a significant impact on both their attitudes and their knowledge can be accomplished. Further study must delineate the impact this curriculum has on behavior and the longevity of the impact. These preliminary results indicate that health education as a technique for violence prevention should be studied further. YOUTH AT RISK PROGRAM (28) The Youth at Risk Program of the Breakthrough Foundation is a San Francisco based program for teenagers who are in serious trouble. The program has two components, a 10-Day intensive experience and a six to twelve month follow-up. Their clients are described as those youth who are failing in school or have been expelled, who are dangerously violent, who break the law, who fight with their families and who are strung out on drugs or alcohol. The 10-Day Course is described as an intensive, rigorous experience at an isolated location. The participants include staff and youth from community agencies. The staff from the agencies are responsible for the follow-up program. The 236 coures is designed to have the participants appreciate the value of integrity, mutual trust and communication, using classroom sessions, group processes, physical exercises and rope courses. "The participants must continually break- through the limitations they have imposed on themselves and the judgements and attitudes they have formed about each other." The follow-up program is managed by each sponsoring agency. It involves ongoing counseling, life skills develop- ment and education. There is one staff person from the sponsoring agencies for each group of ten youth from the agency. This staff person is responsible for the follow-up training of the participants and is referred to as a coach to help the youth stay straight. Each 10-Day Course involves eighty participants. The total cost for the 10-day course and the follow-up counseling is $3000/per participant. The first 10-Day Course was held in 1982. The courses have been run yearly since then in the San Francisco area and recently in the Boston and New York areas. An evaluation of the Youth at Risk Program by an outside agency was released in March, 1984. The evaluation compared the recidivism rate for forty-nine 1982 course participants to a matched group of other youth on probation. The subse- quent incidence recidivism rate was 34.7% for the program participants and 55.1% for the control youth. The serious offense recidivism rates were 18.4% for the program partici- pants and 40.8% for the control. This reduction in recidi- vism rates increased over time after the 10-Day Course and was statistically significant only after seven months. The report concludes that the Youth at Risk Porgram is responsi- ble for these differences in recidivism rates. PEER DYNAMICS (29) Peer Dynamics, a school based program, 1977-1980, was spon- sored by the Nebraska Commission on Drugs, Lincoln, Nebraska. The program was designed to reduce the incidence of the destructive risk-taking behaviors of juvenile delinquency and drug-alcohol abuse among school age youth. Recognizing the importance of peer pressure, Peer Dynamics instituted peer education and counseling systems in fifty-six Nebraska public schools. The program trained and supervised students who participated in a group interaction plan with other students to develop self-esteem and better communication skills. Evaluation of the program was done with pre and post testing for attitudes and a survey form for vandalism activi- ty, grades, dropouts, and contact with law enforcement agen- cies. The first year data revealed an overall positive atti- tude change among peer group members due to participation in the program. Data comparing the program participants to other students showed a noticeable drop in discipline referrals among students. The final evaluation noted that the program 237 affected each sex equally and affected all grade levels. The greatest attitude changes were noted in grades eight, ten and eleven. A control group showed no significant change in their attitudes toward themselves or others. TAKING CHARGE (30) Taking Charge is a school based program for drug abuse and dropout prevention, which has a model applicable to violence prevention. The program is in the Mesa, Arizona Public Schools and developed as a response to strict disci- plinary drug policy instituted for grades seven through twelve. The advent of the new strict policy resulted in a rise in the suspension and dropout rate. The program began in the 1983-84 school year with the goal of keeping students in schools. Once suspended for buying or selling drugs, students are given the choice of being out of school for a semester or enrolling in Taking Charge. Those who enroll sign a contract between themselves, their parents and their school promising not to use or posess drugs. If the student does not keep the contract he is then suspended. The program then requires the student to attend at least eight weekly after school education and counseling sessions which are designed to improve self-esteem and teach factual information about drugs. The program has a multi-institutional base including public schools, the City Mental Health Department, and the local law enforcement agencies. Evaluation data are not yet available but those involved fell that its has reduced drug usage and has returned many students to school. The emphasis on self-esteem is appropriate developmentally and also is a fundamental concept to preventing many of the problems of adolescents. STRIKE II (31) Strike II is a court based program linking juvenile justice with health care. The program is a consolidated effort between the Hopkins Adolescent Program of John Hopkins Hospital and the Juvenile Justice System of Baltimore County. Its clients are postajudicatory first time offenders for violent crimes, assault, robbery, arson and breaking and entering. Those who are not institutionalized and are sen- tenced to a standard probationary period are eligible for the program. Participation in the program is a condition of probation. The program staff includes counselors and para- legal staff. One of the counselors is present in court for the trial and the initial contact is made immediately after sentencing. Within 10 days the meeting with the client counselor occurs and the psychosocial evaluation is started. The client's family is included in the evaluation process at least one parent or guardian has to attend the evaluation sessions at the John Hopkins Hospital. A general physical and a one hour meeting with a psychiatrist are scheduled for 238 the client. The parent meets with a social worker to docu- ment the clients history including school problems and family problems. After the initial evaluations the clients are involved in recreation, education, job readiness, ongoing counseling and ongoing medical care if needed. The medical services are provided through the Hopkins Adolescent Program and are a requisite for participation in the program. In addition to the Stirke II services the client continues with the tradi- tional probation services. The recidivism rates for the program clinets were con- siderably lower that those for standard probationers. Strike II reported a 7% recidivism rate, compared to 35% statewide and 65% within the correction institutions. The basic core cost, not including the medical of job readiness services was reported as $100/per client in July, 1984. The five programs presented are examples of interdisci- plinary, multi-institutional efforts to prevent at risk behavior in adolescents. The programs are developmentally appropriate for adolescents. Three of the five are school based programs which utilize the resources of community and health care agencies. School based programs are increasingly more popular. One distinct advantage is access and availabi- lity, as demonstrated by successful school based health care models (32). Another advantage is the near mandate for an interdisciplinary approach once outside service providers and school service providers operate in proximity of one another. Adolescents who are not in school will not be served by these efforts. The other two of the five programmatic examples are court and community agency based programs. These offer services to adolescents outside of the school. Such adolescent are often in greater need which explains the intensity of the Youth at Risk and Strike II experiences. The problem of interpersonal or acquaintance violence is a severe one. Particulary for young black men in lower socioeconomic neighborhoods. Effective prevention efforts must be culturally and developmentally appropriate. Inter- disciplinary and multi-institutional programs are models for future efforts. Programmatic examples with moderate suc- cesses must be expanded and evaluated further if the 1990 objective is to be met. 239 REFERENCES 1. Centers for Disease Control: "Violent Deaths Among Per- sons 15-24 Years of Age - United States, 1970-78." Morbidity and Mortality Weekly EfifiojJi 1983;32:35:453-457. 2. Alcohol, Drug Abuse, and Mental Health Administration. "Symposium on Homicide Among Black Males." Public Health ReportsP November-December 1980, Vol 95:6 p 549. 3. Centers for Disease control. "Homicide." Morbidity and Mortality Weekly Report. November 12, 1982, vol 31:44, p 594. 4. Barancik JI. "Northeastern Ohio Trauma Study: I. Magni- tude of the Problem." American Journal _£ Public Health July 1983, Vol 73:7 pp 746-51. 5. Gallagher SS et al. "A Strategy for the Reduction of Childhood injuries in Massachusetts: SCIPP." The. Ufitt England Journal __L Medicine. 1982 vol 307 pp 1015-19. Assault data unpublished. 6. Johnson CJ et al. "Student Injuries Due to Aggressive Behavior in the Seattle Public Schools During the School Year 1969-70." American Jflmnal -1 Public Health 1974 vol 64 p904. 7. Baker SP, Dietz PE. "Injury Prevention - Interpersonal violence. Healthy Peopie-The Surgeon General's Report Or Health Promotion and Disease Prevention. Background Papers? U.S. Department of Health Education and Welfare Publication No. 79-55071A 1979 pp71-74. 8. The Boston Commission on Safe Public Schools. Making Qux Schools £_££. fox Learning. November, 1983 ppl2-16. 9. Prothrow-Stith D. Unpublished data from the Boston Youth Program Violence Prevention Curriculum Evaluation. Boston City Hospital, Boston, MA. 10. Centerwall B. "Race, Socioeconomic Status and Domestic Homicide, Atalnta 1971-72." American Journal __l Public Health. 1984 vol 74: 1813-15. 11. Joint Center for Political Studies. "A Fighting Chance for Black Youth." Focus September, 1985 vol 13:9 p4. 12. U. S. Public Health Service. "Promoting Health/Preven- ting Disease: Public Health Service Implementation Plans for Attaining the Objectives for the Nation." Public Health Reports. Supplement to the September-October 1983 Issue, pl67. 13. Centers for Disease Control. "Homicide." Morbidity and 240 Mortality Weekly Reportr November 12, 1982, vol 31:44, p 594. 14. Barancik J. Op cit, 4. 15. Meredith N. "The Murder Epidemic." Science 84, Decem- ber, 1984 p42. 16. CDC. Op cit, 13. 17. Boldrick LN. "A Commonality Analysis: Assessing Unique Effects of Person and Environment Variables on School Per- formance of Disruptive Adolescents." National Institutes of Mental Health (DHEW), Center for Studies of Crime and Delin- quency, 1982. 18. Little JW. "Delinquency Prevention Selective Organiza- tional Change in the School." Department of Justice, Office of Juvenile Justice and Delinquency Prevention, 1981. 19. Dennis RE. "Homicide Among Black Males: Social Costs to Families and Communities." Public Health ReportsP November- December 1980, vol 95:6 p556. 20. Zuckerman D, Zuckerman B. "Television's Impact on Chil- dren." Accepted for publication in Pediatrics. Child Deve- lopment Unit, Boston City Hospital. 21. Williams TM. "Differential impact of TV on Children: A Natural Experiment in Communities with and without TV. Paper presented at the meeting of the International Society for Research on Aggression, Washington, D.C, 1978. 22. Jessor R, Jessor SL. Problem Behavior and Psychosocial Developments A Longitudinal Study _£ Youth. New York: Academic Press, 1977. 23. Erickson E. Identity. v_____ _j__ Crisis. New York: W.W. Norton, 1968. 24. Akbar N. "Homicide Among Black Males: Causal Factors." Public Health Reports. November-December 1980, vol 95:6 p549. 25. Worrell J edt. Psychological PevelOPment In the. Elemen- tary Years Academic Press, 1982; chpt. 3, "The Development and Regulation of aggression in Young Children, pp97-149 (200 references). 26. THE BOSTON YOUTH PROGRAM. Howard Spivak, M.D., director. Boston City Hospital, 818 Harrison Ave., Boston, MA. (617) 424-5196. 27. Prothrow-Stith D, McArdle P, Lamb GA. "The Value of Violence Prevention Health Education in an Inner City School." Boston Youth Program, unpublished data. 241 28. THE YOUTH AT RISK PROGRAM. The Breakthrough Foundation, 1990 Lombard Street, San Francisco, CA 94123. (415) 563-2100. 29. Cooper C. PEER DYNAMICS. Final Evaluation Report. 1979-1980. Nebraska State Commission on Drugs, Lincoln, Nebraska State Department of Health, Lincoln. (BBB18770). 30. Olson-Raymer G. "Community Leaders Promote School Safe- ty." School Safety: The National School Safety Center News Journal, Fall 1984. (TAKING CHARGE Byron McKenna, Director of Guidance Services, Mesa Public School District, 549 Stap- ling Drive, Mesa, Arizona 85203. (602) 898-7938. 31. STRIKE II. Hopkins Adolescent Program, Alain Joffe, M.D. John Hopkins Hospital, Park Building, Room 207, Baltimore, MD 21205. (301) 9556143. 32. School-Based Health Services: Three Approaches. Work- shop for the National Meeting of the Program to Consolidate Health Service for High-Risk Young People. The Robert Wood Johnson Foundation. Princeton, New Jersey. October 3, 1985. 242 Interpersonal Violence: A Comprehensive Model in a Hospital Setting From Policy to Program Karil S. Klingbeil, M.S.W., A.C.S.W. Assistant Professor School of Social Work University of Washington Assistant Administrator Director of Social Work Harborview Medical Center Seattle, Washington INTERPERSONAL VIOLENCE: A COMPREHENSIVE MODEL IN A HOSPITAL SETTING FROM POLICY TO PROGRAM BACKGROUND: Over the past several years, there has been increasing focus on Family Violence; the issues, the characteristics, the components, etiology and intervention/prevention strategies. Although a significant dilemma which has afflicted our society for centuries, this focus on violence comes at a time when we have evidenced dramatic changes in the health of our citizenry. Many successful advances in illness and communicable disease are well known. The attention to traumatic injuries, then, and the development of major emergency facilities and trauma centers across the nation have literally forced health care providers to deal with all types of catastrophic injuries. Included are trauma injuries from interpersonal violence; both intra-family violence and extra-family violence. Gunshot wounds, knifings, physical beatings from other "lethal" weapons, sexual assaults, elderly abuse and the psychological aftermath plague the provider. The picture that has emerged is a frightening one. Clearly, family violence is a major public health concern and requires a community response. As with most health care issues, the clinical demands and frustrations proceeded the scientific explorations and family violence has followed this pattern. Particularly evident over the years has been the frustration of emergency room personnel to the fallout of family violence as physical injuries from this source have dramatically escalated and police, medics and others have brought victims by the thousands for life-threatening care. While many have literally been saved from "death's door," others have not been so lucky as statistics and media headlines have blazened the grueling stories. Even those "saved" often return to the precarious environment from whence they came, only to repeat their journeys to the emergency facilities much to the disgust and ongoing frustration of the health care, criminal justice and social service systems. Further, while emergency medicine has joined other well known medical specialties, emergency room personnel are still ill-prepared to deal with the emotional and psychological impact of family violence. Emergency rooms also function as social service agencies after 5:00 p.m. and on weekends when most agencies are closed and victims of violence have gravitated to emer- gency rooms for both psychological as well as physical attention. Whether emergency room staff are prepared or 245 not, they must deal on a daily basis, 24 hours a day, 7 days a week with all aspects of interpersonal violence. In spite of the frequent incidents, emergency room personnel as other professionals often ignore, minimize, scapegoat, and disbelieve the many manifestations of family violence; therefore the need for education training and sensitive policies and programs become paramount. It was from this clinical morass coupled with an interest and commitment to help as well as to prevent, that the Harborview Medical Center project on Interpersonal Violence developed. The results of violence are most frequently treated in hospital emergency rooms. This reality provides the emergency room setting with a unique opportunity to identify and intervene with high risk individuals if there is a program aimed at identification, assessment and treatment of interpersonal violence. The emergency room provides access to a population that often is too frightened or ashamed to seek assistance from traditional social work agencies, rather patients seek the anonymity of a large, busy often hoped for non-personal health facility. (Clement, J., unpublished paper, 1985). Harborview Medical Center, founded in 1877, is a 340-bed tertiary care teaching hospital affiliated with the University of Washington in Seattle. From its inception, Harborview has served the indigent medically ill of King County. It's priority populations announced by the Board of Trustees in 1984 continue to focus on its initial goal -- to serve persons incarcerated in the King County jail; mentally ill patients particularly those treated involun- tarily; persons with sexually transmitted diseases; sub- stance abusers; indigent patients without third party coverage; non-English speaking poor; trauma patients; burn patients; and those requiring specialized emergency care (victims/perpetrators of violence). Harborview Medical Center is also the major emergency facility in Seattle, King County and the Pacific Northwest. Most medical, psychiatric and psychosocial emergencies are brought to Harborview*s Emergency Trauma Center (ETC). Specific county commitments with the Division of Human Services including the Involuntary Treatment System (ITS) and the Division of Alcohol Services (DAS) bring the acute disturbed psychiatric patient and the excessively intoxicated patient for care. In addition, because Harborview is a designated regional trauma center for the Pacific Northwest, it provides care to the majority of multiply injured patients in Seattle and surrounding environs including the State of Washington, Alaska, Montana and Idaho, referred to as the 246 co O) Cn rO $- O) 73 > O) rO C O-Q r- E uj O OVr- 3 3 O $- >> jc +-> ■M C 3 E O o o •r- M-73 4- C • r« rO C S- O +J >»-r- e -Mi— i— *r- O) +-> $- o rO O i— i—• O 3 S - CX. < O O 3 ^ Q- O) T3 rd o E 0) i— n o 73 r0 i- O U O Cd 4-» -r-3 X C >> CO Q) t- i— O E Ci— ro C rO O •«- O) $- 0) o cn Cn-C 73 O $- O +-> C 00 0) $- ro E Q. C O) LU Q. 0) I— O) O T- -C i— • M- M O) CO «r- ♦ > "O +-> "O M- 0) 0) C O) o 73 0) 0) +■> C 73 r0 S- O -i- T- O ■MO) -MM jC 0) •■- O 73 +J S- C O) C O) -r- $- r0 O 5 ••- +-> CO C CO o O O) rO «i- i— •i- > nj +J t- $_ C O rO CO rO 0) ••- i— C >-o 3 O E S- 0) O. O. O) 0) 21 O CO r— 4-> a O T3 Q. $- O C O 73 $_ rOM- 4_ c O «r- O M OVr- $- O O i-h S- CO O) 4-» 0) 0) C -C ro +-> cn 0) 4- X) O E co O 73 O O) Q£ 0) c >» O 0) c +-> 0) rO Cn 3 $-r- 0) rO E > UJ 0) a> o JC M co C ro •r- jS E co rO r— J- ro cn o O cn $- a. a) JC -* M $- OM- 2 O i— O) (O C •r-O O o oo c O 4-> E "~ Q. rO $- -M E ■O rrj ~ aj +-> 4-> 0) 3 i— $- rO >» CO ^ <« as > O M O •r-4- C 3i— -O O -O +-> O rO O O 3 CO ■MO) S- -Q " J- - Ci. CO rO >> _ >> O CO X O) CU -Q S- CO O) M 73 ro jC •*- c cn o U ro -i- ••- ■r- rO 3 C M O) $- CO 0) co -M -M >» 0) O > a> i*»« jc cn 0) E O) CU CO CO o s- rO CO »r-+->MOEO •r- jC Or— Q-5 O •«- rO CO > -M rO E "O "O <0 $- 0) O) >» rO C -M cu C jc -C oo 4- ••— rO C CO >r- M MO JC O CU Cn n3 >»M t- CH Q.-M JC C JC 0) rO M C«r- 3+Jt-M CO E i- <0 •r- JC O +-> (O OS- 73COJCO)CMOCO O •r- s_ cn c O) c •»- »>i— >a)3T-corOrO-MO)Q. 0730M-0)0) t-coX ^_rOS.O)S.EcO-M30) Q. O) JC 73 O. rO <0 i— 4-> CO CU i— CU C O0)t-73$-OS- •r- cno)MJCjc a> a jc •«- CO MI—J-C03 73«r-CC0 OrOOCT 0) 73 O) <0 >» r— O-r— CU M •«- i— 3JO «Q. rOi- CO > O CO X S- 0) O -r- r— -»-> •<- 0) O COM >S->rOcoco *• jc CO. O3O0)0)-MCn •r- i— OVO C JO S- C 'f- 73 (O CO 3 0) E >,CCO)MrOON-i- r- rO OJC 0'r-4-»«r-M C jC C0r-C73 0)*00 cn o> s- 73 co o.t- •r- O 0) -M >» O) M jCCQ. «Cl-0)r0a>r0 CU S- CU CU CO C JC S- CD ,— CU +-> •>- E 4->MO) l_O-Mr0M-r->»r0 O CU -r- C-M rO S- rOjCM- rO 5 >•»- CO O.O.EM Or— JC 0) r— M JC O) «r- M TO 2* O 0) O) E >>-0*M r- 3 «r- -Or-i— 5 0) > S- M cu a. CO C JC O r— O) rO •!- CO 0) rO O O -r- co M E S- rO Q) ^^M- O) 3 JC (UJJr- +-> ».-r- E *0 CO S- Q) > ► E C3 E UJ S. 0) M > rO rO CnM CU C O JC •r- -r- J— S- > 0) M M • M CO r— rO 0) O JO o u C O >> o •r- $- E M rO rO • .a a> < o C M O) r- r- 3 O 73 •r-< > O) r— JC rO M c o * CO r— s- o 0) o a. o s- M 0) o M S. CO. 0) <4- CO O 3 JO CO <£ O O T-T3 0) -M r- 0) M O «r- JC O) «r- JC I—-o >o 73 C 0) 0) JC > co &- •r- O) r- -M J3 C rO ••- M CO 73 0) C rO CO r- M O C O 0) O E M co O CO S- 0) a. co co r— rO rO O * •r- c C O 0) O c cn s- M C Ol S- -r- rO 73 O) O. 3 CO r- 3 »UJ3 cn c < c ••- •r- $- S- CO O) Q) r— 73 M Or- M O UJ rO O JO M O) OJC Q) S- M CO Q_ 3 73 O M C O-r— rO CO 3 rO <~-« *> CO O) cn CO r— C < *0 •r- E $-r-0) a> > CO (/> Q_ CU O O) •r- JC > I— S- 0) CO • CO co E O) •«- CO M CO O Q) ••- s- > ■a 73 0) rO 73 r- O O t- O E O O M JC O . $-73 a. c rO c O 0) •r- 73 M ••- O O ro «r- 0) 3 al co o o co 3 a. s- r0 O JC M M C •r- 0) Sr- O CO -r- M > C 0) c •r- O) M CU rO J3 O. 0) O > M rO JC co a> o o JC •r- 3 > $- E O) CO CO •!- CO o 0) JC T3 O •r- O > r- O ro $- a. $- o O CO O CO o cu M C Or- $-i— O 0) JC M 3 rO $- CO O) E >)+J CL-r- JQ •-• O M 0) O JC 73 73 «r- r-O) • C > M $- rO id ai a) s M -^ 73 CO • O $- rO O Jjri «r- O O JC CO 73 3 $- M •r- JQ $- COr- >> •r- rO 0) r- M ••- JQ C rO CO O O s CO O ro 0) CO t- O) 0)0 $-73 $- O O) 0) 3 ro $- JC M r- O.M «r- O T> $- C C C >» O •«- a) on •r- O) 0) M s jc ro O C M -M JC C O) C M 0) > 0) M r- $- E <0 $- O CU CU JC CU •r- 4-> > M JC > C r- M •r- O Q) rO $- > > $- O O) C «r- CJ- JC t- 4-> S M ro M i— $- $- 3 rO M- O O) O •r- OM- Q. M E 0) C cn rO ••- r— 0) C »r- 3 M •»- $- CO $- O C 0) C s Q. C M •»- •r- «^ ro rO cu cn s- E JC 0) O 0) CO ■M JO s $- rO •r-73 E C rO ro $-*4- O C *»- O • JC O -r- "-3 O M rO CO C ^ O E 0) M $- O) > C O-r- $- Q) O.J3 O) E <0 O M O) $- C r- i— Q--r- C_J rO «w-r-r- >, M r- r- •r- ro $- . C rO co *r- if- O) E ° , •r- CO $- M T- C O O JC O c»- $- O c O o •r- 3 c at E r— •r- >> rd rO o $- •r- ^ c E cn o 0) M o O ». cn c $- oo r— $- 0) o. E •r- O) E 0) E M j^ OJJUJ ro $- M cn = O) O c $- 3t: •r~ 73 H CU r^ C r— $-^ t^ t: rO $- C •r— 0) 73 •—i rO O^- c oo O 0) ro cn ^ OO $- i—l M c > M O) o c CO • CU CU rO *-^M O LO O) CO 73 OO 73 o) cn a) •r- r-1 O JC CJ- ■*-> 4^ M $- >» CO c a> 0) m 0) Ol-S<£ 0) 73 rO CO •r- O. CO •r- 73 >»T3 C • r- a> >> ro c $- JC JC o ro co Q. 0) t- O) ••- O O M r- r— CO C C o ja o a) 0) 3r- r- > a> q.t- o a) $- C JC -r- $- rO 3 a. > Q. 0) CO JC •!- M $- C 0) O 73 (O C O) o $- M O) ro JC E M $- o *> *M E c o •r- O a: 'ro >> c o o c •r- Q) m cn •r- $_ 73 0) T3 E rO Ul •» $- E M 0) co C JO CO Q) E O) e cu oo O) > CO r- o< ro O $- O 0) -; . #» -*: E $- o o O 2 QC r^ >» « o •r— c O 0) O cnoo $- 0) 73 >E c Ul ro 0) JC JC -M M r— ro C 0) •r- 3T OT3JJ ,r" >*? MOO. O OQ 0) $- M 73 O) o> c cn Q ro C .. . z. c ^ o. o oo § M ,-- • C •r— CO CU O) a> > -Q •r— $- cn r^ 0) c •^ M •r- E C r— ro (H b^ u_ ~ CO -M Ec cu co $- o Of- > 0) O-i— O C r— O ~ rO C 4- «r- O >> O fl) rO •!-> *r- 0)$-CCn>> • O O OO -M S-rO +-> C rO +-> 73 OrOr- r— «r- 3 4- rO Q. oo o E oo o O >»•»- r— COr— -MCr— >»OC0C0 CO •!-•!- EO-MO)rO'i-0)COr—+J co 4- $- ro 73 oo COO'r- J O) •>- t- t- w OJ 0) OO) CO 0) o> a> c +-> •> >-m oo e o +-> o. co 73 jo a) > cn-M a)*r-Ecoo)cccos-a)(0 cos-a)c jc •r-$-0)r— COrO$-0)0)0)0)C-r- fl) fl) +-> rd fl) M C00)T3O'i-O O. > -M O -M $- T-+J3 JC CE *r- $-co 0)4-0) S-CL $-C-M0)-M$- a> a> - > cu cn>>s-occa)o o) •—i -r- +j o JC C r— >> C S- CX. 'i-jCS- co +j ro co 4- 0) r— O 4- «r- O -r- *0 P 01 -MCX. CC0S--r- $-r0«r-OJCC-MT3>»3jCC0OQ. (OOCrO CO • Q.-M -M 3EO)-MCS-ja-MC (0 OO.C0 Q.r- E •«-•!- co Cn O) O ro 0"0 C r— O fl) 5 fl> fl> OO.CE "S-COOEr— C0-I-C73 «r- rd C0O-MT3 O CO Cn $- i— O) 0)-r-r— t-4-> O) corO$-r— E"O-MCC0>73r— ••-> +-> r— CU O fO 00 O O t— +-> fl) CO E C +■> CO •!- C C CU fO 00 00 O) E CO 00 JC CO cu >S-C O O)-MC0C0JCO)C CO ■M4-JCjQT-S-73(Oa)0)r- o> :* $- ai 0)0) Oh— • r073 jc O S- oo E a. jc jC $- cnjc $-000 i— 0)-r-roo04->$-0 ro -Mr— O. OO) Q. C -w- +J rO r— -M rO 3 O. CO $- CO 0)*$-$- O) C • -M JQ $- OO o o ex. c cn •»- C $- $- O- O- $_ r— JC O) >>••- rOa)0)0)-M00300 O • O) E (UOOE-MQ-O-MOO O)C0t-$- $--r-7373 Q.4- O q..i- -r- 4-> CU oo -r- 73 C >> $- -MO. a)-MO)CrOOO ro>jCro»r-Oi— C 0)$- cn<0 CX.rOO.cOO. O- 2fl)OJCCX.Or—r— C04-C$-fl) f073O COO) r— $-0 O) 0)O-MO«r-fl)jC aCi— 4-C0O)JC rj)rO4--MO0rO$-O073*r-C 73TD-M fl) fl) r— «r- -r- M ceo •r-o>a)oo3T- cn e > 0) jc $- •r-O 73$-C73 E EO)r-C0$- C E 0) CO -M 0) 4- 4JC0EC3'f-CC0 i— ••- 3 O C O «r- O 73 -M CO O cd $- ••- cd O r030)r0r-C0CO4- +■> O ••- E _ $- O) rd T3 O JC C CX. oo -r-O rOO)0) orOM encx. 4Jco)0)0-MOE-r- ^ c $- $- .o c J- c » O Cn t-4-CO) -M0)00) COJO ^$-0$-0)$--r- 0)0)0) #r~ §= «=+->*•-> Eoo-r-rdr-S-O) 0)CfO+J +JJCO-M-M O. •r^CCdoOCi— Ocdfl)0_ T3CM<0 C -M C O O 4-> O •r EVIOJJT-r >(. .-r-O0)O •<- fl) fl) 0) f0 I— CO "o)C0rOr- O0)0)r-c000CnO JC r- -o'rp E 0) .r-$-O)O0$-+JCX.0)jB-M<0C$-T3r- 0)-MOOp$-> jCOJCOOO)COQ.Q-OCOOO)3r- £ 'II IT *T « t? -2 il «J> +3 rO CLO) rOOOJC-r- cnO CX.JO rO I— S > O- Q.4- 73 00 ce a. •• s -J O UJ O Q O c ro 73 rjj C JC E <0 M ro $- CO 73 0) 73 OJC C E C O -M 3 O ro $- >) O O CX.S $- Ul fl) >» rO O T3 ■M o| $- C ro O) ro O E -O 0) 4- r- o cn Q. C E C-r- •r- o c •r- $- r— -M 0) 3 fO > 4-r- O CO 3 cn OO O fl) -r- fl) O 4-> JC O $- M 3 rO CO >, $-JO 0) ro JC 0) 73 ■M r- 0) O C C O •I- O) «l- JC M Q.-M O 0) C M CO rO CO »r- 00 0) <0 2 M 0) -M CO $- C $- 3 O) >r- "O E 4- 0) 0) O +-> 0) O rO JC $- 4-> h- O. CO •r- JC • OO CO r— -M O) 0) ro ••- JC JO C ro t- O) O) -M co C r- O •r- 3 ro t- 73 E O O > 3 0) O ••- O $- O 4-> >>i— O. CO «r- r— O $- -r- co 4- O E O •r-O ro M JC oo 4- r- 0) t- 0) Cn 73 3 rO 73 C C C ro «r- C 0) ro M O JC i— C •r- M oo ro O M O) C O 3 C O O +-> •«- i_ 00 00 •r- 3 $- M M $- O 0) C co ro co O. Q) C 0) 0) $- •r-r- $- O 0) 4- C >> O «r- O 3 CO CO r- -r- O $- 0) CO O 73 C C 0) 0) cu cn M T3 M ■4- C 4-> O ro T3 >> C T3 3 O JO 4- CO O O) > CO -r- M M C ro O) i— C O) O $- Q. E C O 0) O 0) CO M 3 0) O JO $1 CO fO »r- > 0) 4- O O C CU c C i— ro co 0) > $- CU 4- co 4- ro M -M C co 0) E4- 0) O M rO C M O CO •!- M O N •r- »r- i— -M O «r- O. $- o o. co * $- M 0) C cn cu C E rO M $- «r- M E 00 E O $- O O O) 00 M O) rO O $- i— O O O fO *r- ••— r- C +■> > O CO O O $- rO co 0) M (O 0) $- -M ro 0) 0) _ O $- _ JO ro oo ro O.T3 M CO M C C O r- E rO O) 3 73 CT o o rO M ~ O i— i C M rO C O CO •r- CO O $- M O) CO «r- rO -r- $- > $- C 0) C m o o. cu co E $- •r- O) O) C 0) M $- •r- +J C ro E ro •-• O "O O < O C JC i— ro M fl) r- r- jC rO co ro ■M rO 0) O JC co +■> JC 0) o cn $- o c •r- 00 OO -r- 3 CO M cr cu oo cu 0) O E Q. $- O rO E $- $- O 73 Q- cn O c o ro >» $- ro $- Q. - ro C CO m cn-r- 0) O) c o cn-r- E $» 73 73 ro 3 3 C $- O JO rO Cn co E O cu CU CU S- $- JC 73 O- M TO O) O) err- u ro cn o. c 3 -r- 0) • O) O -M i— M E $- i— O C >r- JC 3 T- O) M M E> E 0) O 0) > T3 I C C rO O C C O 4- -r- O 4-> c >> 0) o > •r- fl) I— $- o o. Q. >» rO $- rO O E M ••- $- M O. C 0) 73 E $- M ro •r- 5 E O E -M O O O. 0) r— M ro co c O $- ■r- O M ••-> 3 rO M E O) O) M C O) •r- 73 i— $- O rO 0) o JC M ro M oo rO — co CU Sr- ro oo O rO •!- JC r— a. 0) o C CO c o O i— CO •!- $- JC a) ex o. O) 4- > 4- -r- rO M M •«- • oo C CU •r- 4- 0 4- -r- M 0) i— ■o co 4- M ro O C fl) ~ >» CO 73 M CU CU -i- $- > I— Q. $- rO 0) fl) 3 $- co cr Policy Example: Harborview Medical Center (Faculty, Staff, Departments) acknowledges a responsibility in the tertiary, secondary and primary prevention of violence. This includes the detection, assessment, and diagnosis of all aspects of interpersonal violence, the identifica- tion of high risk individuals and groups, and the development of resource networks and/or primary preventative efforts as resources permit. The policy addresses two major areas: A. Intra-Family Violence Family violence includes: 1. Spouse battering 2. Wife battering 3. Marital rape 4. Child abuse 5. Incest (child sexual abuse) 6. Sibling abuse 7. Elderly abuse 8. Abuse of parents by their children B. Extra-Family Violence Non-familv violence includes: All forms of violent acts against another person(s) not related in an intimate situation. Suicides attempts (and homicides) are frequently present in both of the above categories of interpersonal violence and require the use of the Psychiatric Protocol for assessment purposes. The interventions may differ but sensitive and nonjudgemental assessment and diagnosis is imperative. It is important to note that the major difference between category A (Intra-Family) and category B (Extra-Family) is in the definition of the relationship, i.e., violence occurring in the context or absence of intimacy. The appointment of a hospital wide committee on Interpersonal Violence is detailed in Step 3, but could be included as part of the policy formation step. (See Step 3 for further elaboration). 2. Background/Justification Data Following the policy statement detailing both aspects of interpersonal violence, the second step (of a comprehensive program or manual) includes the justifi- cation of the program with appropriate background and 249 substantiating information. This includes information on the specific problem, such as child abuse, wife- battering, suicide attempts, etc. This second second step also requires definition of terminology including the elaboration of different kinds of abuse including the distinction between abuse and battering behavior. Abuses occur in the physical, psychological, sexual and environ- mental contexts. Additionally, there should be a statement of philosophy, principles, and the identifica- tion of high risk individuals or groups in which behavior- al characteristics or descriptions are utilized as the critical identifiers for diagnostic purposes. Thus, step 2 addresses the following key areas: A. problem statement B. definition of terms C. philosophy D. principles: standards of practice E. magnitude of the problem - statistics F. demographics, if applicable G. behavioral characteristics or descriptions H. identification of high risk individuals/groups I. bibliography and references This second step clearly articulates the justification and philosophy for the violence program in the Emergency Room. In otherwords, anyone reading Step 2 would be impacted about the extent of the problem(s), the need for intervention, methods of intervention, etc. Philosophy is critically important to any program but particularly in one where there are varying opinions and prevalent myths that prohibit appropriate diagnosis and intervention. A comprehensive philosophy should include a statement of stand on non-violence as a way of life - that violence is not justified in any relationship except in physical self defense. Additionally, it should address the continuum of violence or the interconnected components from child abuse to elderly abuse. This approach clearly addresses prevention and lays the groundwork for education of client and professional. A definition of violence can be relatively simple and uncomplicated: "...family violence is defined as behavioral toward a family member that would evoke legal action if directed toward a stranger." ("Family Violence Principles of Intervention and Prevention," Jean Goodwin, M.D., Hospital and Community Psychiatry. Oct. 1985). Or family violence includes any act of force or coercion 250 against another person without their permission. Additional philosophical statements can be directed at treatment strategies, advocacy, community resource building and networking. 3. Procedures The third step in the model is the development of spe- cific procedures which include the recommended approaches to the various problem areas in interpersonal violence. Obviously the Emergency Room would have quite specific and detailed procedures, namely protocols, pertaining to interpersonal violence while other departments in the hospital might be much less involved. All departments, however, should have written procedures in concert with the overall hospital policy. As an example, a nursing department policy/procedural statement might address necessary staff development and inservice of the nursing person- nel as supportive to the overall hospital violence policy. This 3rd step would include protocols if clinically indicative and if the department is involved in "hands on" tertiary care...Otherwise, a procedure illustrating how the Department policies mesh with hospital policy will suffice. Brochures A sub-step is the development of literature including brochures and pamphlets on community resources as handouts to patients and their families. Copies of regulations and/or the law should also be available and either be an attachment to a specific protocol or referred to in a Department procedure by citation. Both brochures and copies of appropriate legislation might be included in the Admission Packet which is provided patients on admission to a hospital but would definately be available in the Emergency Room or other ambulatory care facilities as handouts. Hospital Committee Appointment of an overall multidisciplinary hospital committee on Interpersonal Violence should be esta- blished. This committee could assure the continued attention to protocols, the referral network, resources and resource allocation including staff 251 time, and budget. The committee should review hospital policy and update as needed. Additionally, the committee could focus on political issues in the community including public-policy issues, state wide as well as local funding issues and environmental trends. Secondary and primary prevention involve activities in violence issues well beyond direct service. Critically important is attention to legislation, advocacy, and testimony regarding proposed legislation in multiple aspects of interpersonal violence issues. A hospital committee can provide leadership through legislative action. An interdisciplinary or multidisciplinary hospital committee could advance staff training on various levels. This could include a range of activities from mention of the program at new employee orientation to "Violence Rounds," a major educational pathway. 4. Protocols The fourth step includes the development of specific clinical protocols for use in the tertiary programs, namely the emergency room. The clinical protocols are obtainable from the social work department at Harborview Medical Center. They include: A) Adult Abuse protocol (See Attachment A) B) Child Abuse protocol C) Sexual Assault Protocol D) Incest protocol E) Elder abuse protocol F) Psychiatric Evaluation protocol G) Alcohol protocol H) Grief Reaction protocol Protocols related to the overall hospital policy provide for clarity, commitment, and exemplary non-judgmental patient care. (See Klingbeil, K and Boyd, V. Battered Women and their Families. Springer Publishers 1984). This is especially important when myths about interpersonal violence and lack of professional training are commonplace. They serve to standardize a level of care regardless of didactic or experiencial training. The Protocols should include all pertinent legislation that applies to the crimes of violence, reporting requirements and victims compensation legislation, if applicable. 5. Resource Management Step 5 relates to the importance of establishing a current resource bank or network of community referrals and agencies to which victims, families 252 and perpetrators of violent acts may be referred. This critical link involves the development of new programs within the hospital and/or community as needs arise and are identified. This is an exceedingly important step in extending the boundaries of the hospital into the community and developing an effective "safety net" for patient care. Conversely, the "community" comes into the hospital. Resources can be identified in a number of ways but should include the following: Criminal justice system for reporting purposes and investigative purposes as well as treatment planning. Social service system for victims, perpetrators and children of violence including mental health, alcohol resources and self help groups. Crisis to long-term facilities should be identified (Hot-lines) Advocacy and legislative groups including professional organizations such as AMA, APA APsyA, NASW, ANA and state-wide organizations. Religion community Welfare agencies Health care system Educational community 6. Organizational Component The sixth step concerns organizational issues of the Interpersonal Violence Program. This topic is exceedingly important and could constitute a separate paper in itself. The reader is referred, as a reference, to the chapter in Health and Social Work written by Clement and Klingbeil entitled "Social Work in the Emergency Room." Specific areas elaborated are as follows: A. Population served B. Practice and Standards by Discipline C. Supervision - Peer Review Leadership D. Knowledge and Skills E. Protocols - Clinical Aspects F. Program Development G. Administrative Structure and Staffing H. Demographics and Trends for the Future (Community) I. Budget 253 While this resource chapter is illustrative of a broader program in an emergency room, it provides a comprehensive framework for organizing and establishing a violence program as well. EXTRA-FAMILY The initial section of the integrating paper focused VIOLENCE COMPONENT primarily on intra-family violence issues. The second section addresses the extra-family violence component utilizing the Family Violence model. The six steps delineated in the integrating paper must be operational and include the identification of "high risk" populations. It is these populations which constitute the extra-family violence portion of the program and for which early identification and intervention may prevent violent episodes, suicides and homicides. The identification of high risk individuals or groups seen in the emergency room is especially important in (secondary) prevention. Clinical impressions tell us that many individuals, particularly those involved in extra-family violence episodes, have previously been seen at hospital and health care settings. To identify these individuals and intervene prior to a violent act, suicide or homicide is, of course, the purpose of secondary prevention. Some of the high-risk categories are, but not limited to the following patient "groups" and descriptions: 1. Psychiatric diagnosis such as depression 2. Alcohol diagnosis including DWI - substance abuse 3. Behaviors, associated with loss, grief, death 4. Isolation 5. Lack of support system 6. Homelessness 7. Previous history of assault/suicidal behavior 8. Unemployment 9. Presence of a weapon or previous arrest for crime, use of weapon, etc. 10. Runaway 11. Single auto accidents 12. Psychosomatic complaints This list essentially constitutes "red-flag" categories and includes antisocial and delinquent behaviors. We are suggesting that these categories are important to early case finding, and early intervention and that these patients most frequently 254 show up in the emergency room. Their identification establishes the need to begin investigative research into screening devices such as violence scales and/or inventories which are particularly useful to an emergency room staff for educative and predictive purposes in clinical care. Such "scales" are in their infancy but existing scales and inventories in the trauma literature could be adapted in interpersonal violence behaviors. Emergency Room staff could apply such scales to individual patients and hopefully begin to predict the level of lethality for further violent episodes. 1. Policy Statement Policy statements should be inclusive of extra-family violence. Definitions, justifications and philosophy should be sufficiently general to include all components of interpersonal violence; i.e. intra-family violence as well as extra-family violence components. 2. Background/Justification Data Same (See Model) 3. Procedures Same (See Model) 4. Protocols Specific inclusions of protocols dealing with psychiatric diagnosis especially depression, substance abuse protocols and grief reaction protocols are essential here. Early intervention involves early screening and detection to the often hidden etiologies of family violence. 5. Resource Management Same. However, this step in the area of extra-family violence requires considerable finesse and politiking to encourage, indeed demand, that community resources respond to the early identification issues in Family Violence. Many mental health agencies are offering services to only the states priority clients who are CO fl) ■o < c rO ■M fl) "O cd JC C C ■m cn CO 0) r— M >> cd cd c c 3 73 JC ■r- $- JC fl) T- O Q- C JC oo rd +J OO E M JC -r- O rd O) •r- 73 fl) E 73 00 -M o s- > * c c M •r- $- 3 CO oo 0) •r- Cd 0) oo cd fl) c cd cd $- 0) O fl) CO JC > > E JC O cd JO O.JC i— C CO 2 fl) o c c $- ■M O •r- CO r— 73 O CO fl) oo o fl) • C cd ro ro CO -M • $-1—1— •r- •«— O fl) O $- C CO M r— C OO O cd o JC -M cd o 0) ex. *»r- o CU CU M >)f) -r- 1— C X c 73 fl) -M 0) C O. O E $- O «r- > O) fl) fl) fl) O 4- ■M O CO O CO 3 73 C Q. > r— 00 C O CO t- O +■» OO 73 Cd CU 00 4- • $- >> o 3 fl) ■M JC O fl) fl) 0) cn o o 73 fl) r— -r- O i— JC • fl) ro $- CO O JC $- JC fl) -M JO > O O O CO JC O rd O. CO O CO 0) 0) JO C cd 4- -r- -r- CU 4-» -r- Cd $- E c co $- -r- jO i— > JC 3 4- C i— O.T3 fl) *r- Cd 3 O cd fl) 3 00 CO -r- -r- nj •*> 0) JC +->>>$- C .o >» CO ■M 0) -M O C -* > $- CO Q. • 00 i— CX. O r- 0) •r— C M C CU •i- o $- o •r- E oo •r- $- CO ■M «r- $- fl) cd o> o C «r- O E fl) cd $- fl) 73 cd -M $- co E 3 r— E $-73 C •r- +J S JC '$- fl) T- fl) o CU 3 Cd -M rd ■M +J -r- fl) I— O M 0) M enje $- >> , cd ex. E 4- cd O •r— 00 r— CJ CU CU > o +-> o r- $- 4- $- l r— •r~ E 3 0) O M C ro c $- o cn fl) O fl) C Cd O c E 1— JC T- 4- fl) JC -r- Ol. 0) ■M 4- C -t-> O $- C •r— O !—+->> OT3r- 0) M 3 •r- C •r-+J CO r— O •^ cd >» co 2 jc cd O >> •^ ■M X E o $-4- fl) 0) $- $- E 0) M O cd -M 73 C fl) O r— $- O O co jc $- M rd U_ ■^ C4- fl) C 0) cd Q) 4- 3 M Q) C $- O 0) T3 C rd O ■M 4- JC c O. CO 4- 3 cn mo C 3 ■M O O •M o rdr- E fl) r- O) O o • c rd -M C C cd -r- •^ O. O) -r- JC -r- $- o s- oo c cu $- O O CO $- co M 73 +j +j rd ex. o o r- >» o •i- «r- ~-r- -M fl) ro 0)0 0 ■M 73 fl) •r- -i- o r- Q.-M -M $- co cd cn N > E T- CO fl) C JC 0) C -M •!- fl) ex. c rd 0) O t- $- •r— •r- > •r- 73 cd M O t- C > $- O fl) r- JC C JC a) c ■M 0) C r— 0) O) > rd ■m cn o E ro fl) cd > 4- O. >> C c fl) O -M >> > >> fl) O $- cd >>jo O E $- -r- O fl) r- O ■r— r— M i— fl) C $- CO 3 -r- CO 3 $- rd cn oo M r- t- O -M CO t- co cd O. 0) u. 73 ex, on O OO fl) C M 00 cd M co r- C E M CU C O C r— > Q) 73 CO C JC CU 1— •!- fl) CO •r- fl) 4- • •r- fl) E cd 4- > M >»M 73 • to This compr treat cruci speci inter ■r— O. CO O JC r— CO 3 CO •r- O-i— $-E +j O M CO 3 C X 4- O «r- Q) >- oo 00 O-r- C fl) C C I M JC M 0) JQ *r- 73 $_ O fl) -M CO > d) O -r +J TJ JC 4-> r— -r- COr— r— .r-4-» coccn-MOcdcn -r- i— rd73 4-MC>>$-CdCO0 C -r- >> t-O) +JC •r- C O) $- +-> •!-•!- 73 Ofl)r— M 3 JC CCd m fl> > cd oo j* ex. cucoojo-m •—• -m cu C E $- E 001—4-0.0)00 C co EO) CU 00 CU -r- '1-0)4-00 730)0) 73$- 30 73 CO +-> $- • $- JC cd r— 73 JC i— 3JO • C O $- -r- •r- fl) C CL C I -MO) 4->3CT 0034- -M > co «r- O -M 73 00 > >»t- O fl) -M $-4- oo $- 0)CO 73 -r- cd C fl) r— j5 JC $- CO O E CO) JOCd>>C+->iCd4-73-M 004-3 Mr— Cd «r-cO r-edcoo o ccn E ord$- co$- o)+j « fl) fl) c c c cd c cn cu i— r0«r-c0+J>CCnCJO3T-OOC0 4-00 JOcO O$-0)C$-0)CO CTJ* ••- -r- fl) -r- $- m- +J CU 0) «r- *r- -r- O M 3 i— MO. T30 C0Cd73EM-M$-+J+J$-OCCdC0 t^ ••- r— O Q.-r- COOCCdcdcd 4- jS fl) i— OO -1-730) 3O0 3jCCdC0-r-O-OO73 4->+J3-r- OT3> O OO fl) 3 fl) JC fl) M Q. Ccd-r- 0>» S_ >,M CO >» >» «»T3 0)OZ:<£0>) cd CO* 4-> cn oo c oo i—4- cnuj c-r- Q.+J c fl) c c >>••- Q.fl)Ma>73oo-MO$-o enje ■*->•*-> a) a> 0.1— o. 0 1— CO fl) +■> CO -r- OVr- CdOO C O073E3 O OOOOOCO O-r-+JO-M-M-r-fl)0) CO-r-OCL fl) •r-CdcdO0fl)Mr— cd$-cdcOr— OT3 0)COOO -MC > fl) > 1— O.I— O- C -r- $- C Q. Cd -r- JCJC$--r-734-3 3>>E-r-4- 73000 > >»OM73-M 0) O O-O- CLr— OOOC 73O-r-fl)-r-C0 VO r—30T3cdO-030$-0 O cdJCO $-fl) «n •r-oojacd$-o-MO.Cd 0) M rdoo*>a>Q.a)a>A)ooo cu o jc 0.1— 00 ■<- 4-r— 1— JOE>Or— r— •r-0)O)-M0) C-MOO cd$- IOO -r-fl)CJOr— JCJCJOOOS- r— -r- O rdOOC 73OC0O-M-M >>C0 oo CU CU > r— -r- $-00(00) O $- U. 4-> 00 O cOi— > r— $- 4J>MMO$-0) O) OCOO JC JO 0) >» 0)0. XOO COJOCDC M-r-3Cn$-« -r-73r— S CU S- i- 00 JC r— • ECO—^ $-00 -r- cn O.CL0)C04->+J3r— MO. -r-4-73 CU 00 O E COC fl) 3-r-CO > rd 3 r— (A j^ 0) O. O M CO CO-r- JC JC r— CT OO 3 0) •r-Ofl)0)O=C0 COO. 4- -M MCnO-r-OEM>>4-> JC-r-COCOCO O- 731 OM 3JCCC Oi— CJC O $- O- 0) CU CU (O -r-fl) a)OOJCcncoEJca)OJ4c:c-MO$- co4->73$- i— oo u$-oocd0)ocnoocda)a) jct3 -i-t-cum joi CJCr-a)-r--r-$_-r-C0a)a)CnOC073 JC 3 fl) X 3fl) O -M cd M T3 Cn O.JC fl)$-cordMs cd r—CTCfl) 0.$- O -M CO $- M cd i— M i— JC fl) O C CO JC cdfl)5+JjOO 3 -r-MO> $-C0) O O JC i— $- $-OC4->CMC0 O MrdrO 0)73 $- CO O i— 0)0 4- fl) O O $- O >t CO OJC fl) fl) > 4- O. O M -r- M JC $-jO-r-4-fl) O-r-JC fl)M 000)0 •+■>!— -M CJCr- MCCO CC0OC0 COMrO CrOc073Q)C$-3CO-r- Q) rO O E fl) i— O O. CT-r- M C JC +J +J O -r- i— C -r- $-jC4-»$- -r-M 73 I— CO O) 73 O O 3 fl) > CO • C rd 3 M CO fl) 0.300 i—73rdMCMM73 E E JC C >» 3 0) -r- Cd CO 00 c •r- -M O i— O.T3 M O- E O CO M -r- O -r- O C -r- CO 4- O O. 0) (O 2 EO.C073OI—O X JC JC fl) CO O- 3 JC rO Cn fl) 4-> -M >>JC 4- fl) X CO $- J* C +J+J OO) >»30-r-4- E ~ fl) CC O $-73 rdr- O $-$--r-JC-r-Q)73+JO)fl)r— CO O0)OM r— C >0 $. W (fl 4->0-r-r— OcOOOUIOCU-r-EO) fl) CO 3 3 -r- O) $- JC O. $- > O 26 4- > 73 > O-r- oo fl) -r- _C O i- C 00 CU i— • Mi— 3 JC 3 -r- 00 >»-r- fl) >»73 C OCnfl)r—4-CO'MC • -i- C 4- CO •» fl) O -r- -r- E -r- CO 0) -r-O fl)$-CJQOE70Q)r- OOO •r-rO-r- 3rOOJC-r->>CM>>00 OO) +Jco4-E+->JOOO)r—i— E C i— CU 00 I CO-r-r— 3C$- fl> O O 3 73 rd ~73 M i— OCOOQ) Cn CEC$-73CCO-r-fl) M (OCOO) 3-M0>a>3 Q-> $- M •r- r— fl) O X Cr— O O C JZI 024-0)0 OCOr— COC-r- 4-J +J -r- r— >> O COO) r— l—•> JC$-0)i— COO)C$-$-M Cd • O O > r— >OC0O3 CU a)C0cO4-O)cdO)rOC0 MJO JC -MCUO 73COJC$-r-0 J* cd O $- 73 OC fl)cdfl)C0 r— $_.r-$_O-fl)fl)-r-Cd73Q.O00S- c0O>3O.MJO4J$-i— $--r- cd +-> 3 fl) O CO O. -r-33fl)Mfl)i— CMi— CO rOC73COOM-r-$-i— fl)fl)r— Q)C73cO"OOOJCC$-cOO ECco$-rorOOrOrOoo»—iOOT3 r- cs HOSPITAL ADMINISTRATION POLICIES PROCEDURES EMERGENCY ROOM INTERPERSONAL VIOLENCE PROGRAM INTRA-FAMILY VIOLENCE COMPONENT EXTRA-FAMILY VIOLENCE COMPONENT Policy Program Protocols Network Referrals/Resources Follow-Up Quality Assurance Research Training Prevention Tertiary Secondary Primary Advocacy-Legi siati on Public Policy Diagram: Hospital Emergency Room Violence Program 258 Attachment A /ADULT ABUSE PROTOCOL (Includes Battered Woman Syndrome) Department of Social Work Harborview Medical Center I. Criteria for Social Work Involvement: All adult patients diagnosed as victims of battering; suspect cases of battering and/or unexplainable physical injuries in which battering is to be ruled out. II. Social Work Intervention A. Emergency Measures (Recorded) 1. I.e., crisis management/intervention and medical consultation as necessary. Issues are "hidden" physical injuries due to minimization and/or denial on part of the patient, and batterer, who may frequently accompany patient to the appointment. 2. Provide Following: a. Photograph injuries b. Obtain consents for Release of Information and photography for documentation of physical injuries if patient desires to file charges. c. Provide emergency room forms for filing charges in the criminal justice system. B. Assessment 1. Process (Not recorded) a. Review existing chart and/or pertinent information, b. Consult with staff. c. Interview patient and identify battering diagnosis as soon as possible. (Observe, if possible, prior to interview). d. Contact relevant agencies who have had contact with victim, i.e.: Seattle Police Department, Sheriff's Office, referral source, etc. e. Contact appropriate support system with patient's permission. 2. Information (To be recorded) 259 Adult Abuse Protocol (Includes Battered Woman Syndrome) a. Reason for referral to facility and/or social worker. b. Patient's mental status and behavior. c. Previous history of abuse - physical and psychological. d. Financial/personal resources. e. Patient's perceptions, attitude toward incident and batterer. f. Danger/lethality of situation for patient/ family members C. Plan (Record as appropriate) 1. Patient will be educated re: the significance of danger for the future, i.e., violence is cumulative and increases in intensity, progressive, etc. 2. Options discussed re: returning to same situation, vis a vis shelters, family, friends. 3. Implications for children, including the danger and lethality of situation, will be discussed with patient. 4. Reduce guilt/blame emphasis by focusing on behavioral characteristics regarding victim as well as batterer. 5. Detailed information will be provided to patient on filing charges, the legal process, advocacy and follow-up needs. 6. Patient will receive referral to appropriate community agency, CPS, Law Enforcement, etc. 7. Mobilize resources, including patient's interpersonal support system. 8. Communication will be enhanced with all necessary individuals in order to develop a support system in patient's behalf including staff, family, clergy, criminal justice system personnel. 9. Patient will receive follow-up medical/counseling care. 10. Redefine goals as appropriate. 260 Adult Abuse Protocol (Includes Battered Woman Syndrome) D. Implementation Process (Not recorded) 1. Interviews as appropriate. 2. Consult with staff, i.e., triage nurse, physician, others, including community agencies. 3. Telephone calls. 4. Obtain Release of Information forms. 5. Complete Social Work Intake Report. E. Outcome (To be recorded) 1. Indicate whether each goal under "C" was achieved and if not, why not. III. Criteria for Terminating Social Work Intervention (Record as appropriate) 1. Goals achieved - specify 2. Non-compliance - specify (including AMA) 3. Referral - specify 4. Hospitalization/moved/other - specify 261 MEDICAL AND SOCIAL WORK PROTOCOL FOR EMERGENCY ROOM/WALK-IN-CLINIC TREATMENT CASES OF ADULT ABUSE (Including Family Abuse) MEDICAL 1. Register victim as Alleged Adult Abuse. 2. Triage Nurse will notify social worker that the victim has been registered. 3. After victim has received medical treatment and has been medically cleared, Medicine/Surgery Service will notify the social worker that the patient is available for an interview. 4. If Triage Nurse determines that victim does not require medical attention, victim should be automatically triaged to social worker. 5. All cases of alleged Adult Abuse to be seen by social worker. SOCIAL WORK 1. Interview all alleged Adult Abuse patients 2. Use Abuse folder 3. After victim is medically cleared, interview victim to assess mental status, behavior, personal support systems, and need for additional services. 4. Assessment Categories: A. Subjective Distress B. Behavioral Disturbance -Anxiety/Fear/Anger -Speech Disorganization -Depression -Agitation/Excitement -Social Isolation -Reported Impulsivity -Suicidal Ideation -Interview-Belligerence -Homicidal Ideation -Inappropriate Affect, -Guilt/Blame Appearance, Behavior -Denial of need for -Reported Overt Anger Therapeutic Intervention -Suicidal Behavior or referral -Homicidal Behavior -Minimization C. Chemical D. Societal Role Impairment -Alcohol Abuse -Wage Earner Role Impairment -Drug Abuse -Housekeeper Role Impairment -Student Role Impairment -Mate Role Impairment -Parental Role Impairment 5. Photograph patient for medical chart. Obtain consents for photography only if patient will be filing charges with law enforcement. 262 Medical and Social Work Protocol For Emergency Room/Walk-in-Clinic Treatment Cases of Adult Abuse (Including Family Abuse) Social Work (cont.) 6. Provide patient with HMC form which explains how to file charges with law enforcement agency. 7. Complete Social Service Intake Report for Battered Woman patient. 8. Disposition -Arrange for escort home, or other shelter if needed (Detox, if necessary) -Give victim information and referral to legal and counseling services. -Contact Children's Protective Service, if necessary. 9. Xerox medical workup and social work evaluation for Adult Abuse folder. 10. If patient is seen in Emergency Room, put folder in SAC drawer. If patient is seen in Walk-in-Clinic, put folder in Social Work drawer. 11. If patient wishes social work/counseling follow-up at Harborview Medical Center, please leave note for Emergency Room Social Worker. 263 Interviewing Persons Hospitalized with Interpersonal Violence Related Injuries: A Pilot Study Burnet B. Sumner, M.S. Assistant Clinical Professor Departments of Surgery and Psychiatry University of California, San Francisco School of Medicine Elizabeth R. Mintz, M.S.W., M.P.H. Departments of Surgery and Urology University of California, San Francisco School of Medicine Patricia L. Brown, M.P.H. Department of Surgery University of California, San Francisco School of Medicine CONTENTS I. Description of the problem and background.......269 II. Specific objectives............ • 270 III. Study sample.................270 IV. Methodology.................271 V. Potential confounding variables and design limitations . . . 272 VI. Findings.................. 273 A. Sociodemographics ......... ..... 273 B. Characteristics of the Injury-producing incident .... 282 C. Role of alcohol and drug use..........287 D. History of recent stress............289 E. Recent exposure to and experience with violence .... 293 F. Recent involvement with the criminal justice system . . . 295 G. Health history...............296 H. Personal and family history ...... ..... 298 I. Developmental problems and stressors ........ 303 J. Cost of injuries..............305 VII. Profile of characteristic violence-related-injury patient . . 306 VIII. Discussion.................307 IX. Recommendations................308 267 FIGURES: Results of Interviews Figure 1. Age.............273 Figure 2. Race.............274 Figure 3. Sex.............275 Figure 4. Education...........277 Figure 5. Employment ........... 277 Figure 6. Household income..........278 Figure 7. Housing...........279 Figure 8. Marital status .......... 280 Figure 8A. Parenthood among respondents ...... 281 Figure 9. Married-single .......... 282 Figure 10. Time of injury-producing Incident ..... 282 Figure 11. Location of injury-producing incident . • • • 283 Figure 12. Mode of injury..........284 Figure 13. Nature of injury-producing incident ..... 286 Figure 14. Relationship of assailant ....... 287 Figure 15. Alcohol consumption prior to injury • . • • • 288 Figure 15A. Alcohol ingestion prior to injury documented in medical record.........288 Figure 16. Non-therapeutic drug use prior to injury. . • • 289 Figure 17. History of recent stress among patients with violence-related injuries ....... 291 Figure 17A. Stressors reported by respondents during previous 12 months.........292 Figure 18. Exposure to and experience with verbal confrontation in last year.......293 Figure 19. Exposure to violence in last year.....294 Figure 20. Experience with violence in last year .... 294 Figure 20A. Exposure/experience with physical violence in last year........... Figure 21. Experience with the criminal justice system • . . 295 Figure 22. Respondents treated for previous injuries . . . 296 Figure 23. Respondents with recent health problems .... 297 Figure 23A. Health problems in order of reported frequency . • 297 Figure 24. Was family intact during childhood? .... 298 Figure 25. Substance-abusing parent or family member? . . . 298 Figure 26. Respondents' experience with precipitous deaths of family or friends...... . „ . 299 Figure 27. Mode of discipline during childhood ..... 300 Figure 28. How often slapped or hit by parents or relatives . 300 Figure 29. How often slapped or hit for no good reason . . . 301 Figure 30. How often did parents slap or hit each other? . . 302 Figure 31. How often have you slapped or hit partner? . . . 302 Figure 32. Developmental problems and stressors reported by respondents ........m 304 Figure 33. Cost of initial inpatient hospitalization for persons with injuries resulting from gunshot, stab wounds, and assaults ..... 306 268 I. DESCRIPTION OF THE PROBLEM AND BACKGROUND OF PROJECT Trauma is the foremost public health problem facing our nation today. Injury is now the leading cause of death for Americans aged eight months to forty-four years of age, greater than half the average lifespan. Added to each incident of premature death are at least ten often seriously disabling and always costly non-fatal injuries. The greatest loss of life and function are among our youth and working age groups. Increasing numbers of deaths from injury are due to interpersonal violence. Among young men -- Blacks and Hispanics in particular -- the rate of increase for both fatal and non-fatal injury is escalating rapidly. Data on injury fatalities is collected nationwide. Analysis of the outstanding features of such deaths is possible. There is, however, no comparable information-gathering system for injuries which do not lead to death. Therefore, while we can infer some of the dynamics of injury deaths, we can only speculate about the factors operating in the tenfold number of non-fatal injuries. Collection of basic sociodemographic, physiologic, and behavioral data on all injury admissions to San Francisco General Hospital Medical Center has been one of the principal tasks of the Trauma Research Projects in the Department of Surgery during the past year. Sponsored by a grant from the Robert Wood Johnson Foundation, this data bank has facilitated the study of the spectrum of Injuries requiring hospital treatment. Similar data on injuries treated and released by our emergency room, by other hospitals, as well as those who die before receiving medical treatment will be incorporated into this data bank when resources allow. Consolidation of such information will make feasible population-based analyses of injury in an urban setting. Findings from this survey of admissions to the County Trauma Center indicate that at least th_r ty_-s ix_P.erc.ent of such injuries are a function of interpersonal violence. That such a high proportion of hospital-treated injuries are intentional at some level is an urgent concern to those responsible for their care. The American public, clinicians included, continues to subscribe to an "accident" model of injury causation. In this "accident" paradigm the injury event occurs simply and randomly. Similar to our selective inattention to the results of drunk driving, our ahistorical view of injury events carries over -- except perhaps in extreme Instances -- into the arena of interpersonal violence-related injuries. The tendency to focus on the wound itself, while ignoring the process and circumstances of the injury-producing incident, precludes genuine understanding of the nature of such injuries, meaningful secondary intervention, or well-conceived prevention programs. This pilot study was initiated in order to increase our understanding of the role of interpersonal violence in injuries. While little more than descriptions of outcomes can be obtained in the case of fatal injuries, there is potential for greater 269 explication of non-fatal injuries. Many patients with violence-related injuries have been treated previously for such injuries. It might be hypothesized that such injuries, especially in those persons with severe or repeated injuries, reflect P_rehomic J.dal behavior syndromes. For these groups of patients, it is both expedient and essential to use such incidents as the starting point for the elucidation of these destructive interactions. Trauma, at first glance, seems a straightforwardly physiologic phenomenon. The opportunity to scrutinize the larger picture leaves one without question that, like neoplastic or cardiovascular disease, injuries reveal comparable risk factors a significant number of which are behavioral and situational. II. SPECIFIC OBJECTIVES A thirty-day pilot study of hospital-treated interpersonal violence-related injuries was undertaken for several purposes. The first was to deveiop__methods for obtaining information which would add to our appreciation of the process and circumstances of these kinds of injuries. A second objective was to ascertain the feasibility, of eliciting pertinent information from this population of patients who might be reticent, as well as in pain and in an unfamiliar environment. A third objective was to examine more specifically the d y_nam ic s_of _vio lenc e_be tween IC3Uaintances with an eye to formulating strategies for intervention. Ill. STUDY SAMPLE The interview sample included 32 of the 105 persons with gunshot wounds or stab wounds or injuries resulting from assault who were admitted between 11/15/84 and 12/15/84. While, certainly, some falls, motor vehicle collisions, and burns represent interpersonal violence, these are not included in this account because differentiation of intentional from unintentional motivation among these modes of injury is exceedingly difficult without extensive interviewing of all persons involved in the incident. Because of the pace and timing of admissions and discharges--a significant portion of patients are admitted for operative procedures on Friday or Saturday nights and are discharged prior to the census count at midnight -- random selection of violence-related injury admissions was not realistic for purposes of this pilot study. Patients who were medically stable, that is, transferred from the recovery room or intensive care unit to the wards, were approached for interview in order of their admission. Excluded from the interview study were patients with language barriers, self-inflicted injuries, those discharged, and those hospitalized on the jail unit. While the jailed group was of considerable interest, the administrative and medical-legal procedures required to obtain permission to 270 interview these patients were too complex to have been dealt with in the timeframe allotted for a pilot study. This group will be included in any future interview studies. Since those with less severe injuries were often discharged before we were able to interview them, this discharged group mentioned above probably skewed our interview sample in the direction of the more severely injured. IV. METHODOLOGY While the medical record is an invaluable tool for many purposes, information on injuries is generally limited to a description of the lesion. Very little is included about the patient himself or the circumstances of the incident. The busy surgery resident has little opportunity or inclination to inquire further than for whatever information is essential to keep his patient alive. Unlike other disease processes, the injury is recorded as an isolated event even in those persons for whom there is a history of previous trauma. Because of the limitations of the medical record, our study sought other vehicles for elucidating the injury process. It seemed to us that the starting point for understanding the patient's experience should be the patient himself. We, therefore, chose a structured interview model for our one-month pilot study. This format could potentially include 3M§Dtitatiye as well as qualitative data relevant to the process and content of the injury-producing incident, the context in which the encounter occurred, information about the E£rson himself and about the environment in which he lives. This mode of data gathering would be difficult, it was acknowledged, on a busy trauma service where acutely injured patients are in pain, undergoing medical and nursing procedures, often withdrawing from alcohol or under the influence of analgesics, and being discharged as rapidly as they were admitted. Since a random sampling of the population of injury admissions was not a goal of this feasibility study, efforts were directed simply to interviewing as many of these patients as was possible in the thirty day time frame alloted to the study. Clinical experience and the current literature revealed a number of recurring themes in violent incidents. Because little work has been done in this area, the study design was principally descriptive, rather than explanatory, in nature. Certain characteristics of the patient and his experience seemed particularly relevant, however, and were targeted for exploration. These included: sociodemographic factors, personal and family history, health history and habits, current housing and working conditions, the extent and condition of the patient's support system, exposure to and experience with violence. Other factors whose association to violence were selected for study were history of recent and cumulative stress; personality traits including self-esteem, aggression, and capacity for self-control; history of emotional thwarting; precipitous losses of important persons; general affective state; 271 and socioeconomic and cultural factors. The interview was conceived of as exploratory and devised so as to capture the flavor of a number of areas of the patient's functioning. Data gathered from such a broad spectrum screening instrument could be used to generate hypotheses for further in-depth studies. A departure point for the construction of our interview were sections of Henry Steadman's protocol used for the study of situational factors in violence among ex-mental patients, ex-offenders, and residents of the general community in Albany, New York. Our interview included two psychometric scales, Harrison Gough's Checklist and the Belief in a Just World Scale. Questions were of two types: structured and P_re-coded , and 2E£!l::6ndied questions which would require ex^gos t _fa ct o content analysis. The interview was designed to last no more than one hour. It was audio-taped to improve the flow and for insurance against the loss of affective data as well as valuable detail. The response format provided for administration by non-psychiatrically trained interviewers and the reduction of bias. The interviewing team included psychiatric clinicians, graduate students in the health sciences, and non-clinical persons with some interviewing experience. Efforts were made to achieve a reasonable sexual and racial-cultural balance among the interviewing staff. V. POTENTIAL CONFOUNDING VARIABLES AND DESIGN LIMITATIONS No report of research findings is truly complete without some attention to the confounding variables and limitations of the study. We shall begin with these since they influenced the conduct of the study. The project was conceived, formulated, and executed in a matter of weeks. Time constraints necessarily influenced the refinement of instruments, the sampling schemes, and the extent of data analysis. A second factor of time was that interviewing was concentrated in a four week period which included holidays, university examination periods, and school vacations. If data collection had occurred during a less complex season, the extraordinarily high rate of admissions might also have produced more interviews. Thirdly, the opportunity to pretest our instruments and survey the range of responses would have revealed the extremely low literacy level encountered which necessitated explanation of items by the interviewers and resulted in some inability and occasional refusal to complete the scales. Finally, more extensive interviewer training would have increased the consistency of data gathering. 1 Steadman, H.S. "A Situational Approach to Violence." International Journal of Law and Psychiatry, Volume 5, 1982, pp. 171-86. 272 VI. FINDINGS A. SOCIODEMOGRAPHICS Despite these shortcomings, the principal objectives of the pilot study were achieved. We have a workable set of instruments. We now appreciate what it takes to obtain access to this population of trauma patients. Our interviewers have provided us with an understanding of some of the problems which these persons are facing and a rich flavor of their lives. AGE The burden of injury falls hardest among our late adolescent age group, as well as those who would be expected to be entering and reaching peak performance levels in the world of work. While 47% of the County population is between the ages of 18-44 years, 71% of persons hospitalized for injury at San Francisco General Hospital are in this age range. Among patients hospitalized for gunshot, stab, or assault wounds 83% are youth or young adults. Because injury tends to occur at an earlier age than do the onset of cancer or heart disease, relatively more working years are lost to death and disability from injury than from either of these other diseases. FIGURE 1 S.F. ALL INJURY COUNTY ADMISSIONS PERSONS 47% 71% BETWEEN 18-44 MEAN AGE 38 36 RANGE UK 0-94 VIOLENCE- INTERVIEW RELATED INJURIES SAMPLE 83% 84% 35 34 15-75 18-75 RACE The racial composition of the City and County of San Francisco is White. 58%; Black, 13%; Hispanic, 12%; Asian, 22%; and Native American, 5%. Whites represent only 80% of their numbers in All Injury Admissions (AIA) and only 57% of their number among Violence-related Injury Admissions (VRIA). By contrast, while Blacks account for only 13% of the General Population (G)) they reflect 21% of All Injury Admissions and 44% of Violence-related Injury Admissions. Blacks are 273 overrepresented by 62% among AIA and almost 240% in VRIA. Hispanics are overrepresented in the All Injury Admissions group by 25%. They are not, however, as significantly overrepresented among our thirty-day sample of Violence-related Injury Admissions. While Asians represent 22% of the General Population, they comprise only 1% of All Injury Admissions and 3% of Violence-related Injury admissions. Native American admissions for All Injury Admissions are less than 1/12 of their proportion in the General Population, which is 5%. They are not represented in this 30-day sample of Violence-related Injury Admissions. Like Hispanics, however, Native Americans are frequently misidentified on hospital admissions forms. Likewise, their racial identifications are often mixed. Historically, Native Americans have high rates of traffic and violence-related injuries. <2> The vagaries of the record keeping systems, and bias in our thirty-day sample of violence-related injuries are the best explanation we can make for these discrepancies. FIGURE 2 RACE ALL INJURY COUNTY Al )MISS 1980 TO S.F. WHITE 58% 47% BLACK 13% 21% HISPANIC 12% 15% ASIAN/ 22% 1% NATIVE AMERICAN 5% .4% UNKNOWN 0 16% V.R.I.A. V.R.I.A. 11/15-12/15/84 11/15-12/15/84 (92) (32) 33% 50% 44% 41% 13% 6% 3% 3% 0 0 7% 0 Simpson, S., Reid, R., Baker, S., Taret, S. Injuries among the Hopi Indians, Journal of the American Medical Association, Volume 249, Number 14, 4/8/83, pp. 1873- 1976. 274 SEX The sexes are evenly distributed in the County population. However, 78% of All Injury Admissions and 90% of Violence-related Injury Admissions are male. Possible reasons for this discrepancy include the socialization of males in our society to use physical, rather than verbal., modes of behavior in dealing with the world; the encouragement of males towards adventuresomeness and risk-taking, while females are encouraged to seek security; the participation of males in occupations with high risk for injury; the culture of compulsive masculinity or machismo-- more pronounced in some groups than others—promotes competitiveness among males and violence as a method of problem solving; finally, increased aggressiveness in males may be contributed to by the absence in many families of an adult male role model for the channelling of aggression as well as sanctions against the acting out of aggressive impulses. FIGURE 3 SEX ALL INJURY V.R.I.A. OF V.R.I.A. S.F. ADMISSIONS 11/15-12/15/84 11/15-12/15/84 COUNTY S.F.G.H. (92) (32) MALES 50% 78% 88% 97% FEMALES 50% 22% 12% 3% INJURY DEATHS REPORTED IN SAN FRANCISCO CITY AND COUNTY 11/15 - 12/14/84 All injury deaths in San Francisco City and County are recorded and autopsied. 49 such deaths occurred during the time period of the interview study. The composition of this group of injury fatalities differs considerably from that of the group of injury admissions to San Francisco General Hospital Medical Center. Mean age of the injury fatalities is 49.7 years, over 15 years older than the mean age of the All Violence-related Injury Admissions, which was 35. Range in age was from prenatal to 95 years. Racial composition was also different. Whites accounted for 71% of deaths as compared to 33% of admissions. Blacks represented 16% of deaths and 44% of injury admissions. 275 Hispanics comprised 8% of injury deaths and 13% of injury admissions. Asians were 2% of fatalities and 3% of admissions. There were no Native Americans among this 30-day sample of injury deaths . Sexual composition of deaths and admissions is somewhat more congruent. 80% of deaths are male, while 90% of admissions are male. 20% of deaths are female, while 10% of admissions are female . The reason for these differences may be several. First, only 12 of 49 or 25% of fatalities appear to be interpersonal violence-related. Of these, there were 5 stabbings (42%). Four or 33% were shootings. Three or 25% were assaults. 21 of 49 or 43% were overdoses and suicides. We have combined these categories because in overdose situations, the distinction of suicidal intent from subintentional and unintentional is difficult. 5 or 10% of injury deaths were traffic-related. 7 of the 49 injury deaths or 14% were falls. Others involved burns, asphyxiation, or medical misadventures. Definitive explanations of these differences will emerge only from further examination of these populations. Several possibilities are suggested by the data. With regard to age, the high number of suicides among older people may play an important role. With regard to racial composition, older white men were overly represented in the suicides and this may have influenced the racial composition as well as the average age of the fatality group. Differences in sexual composition may be due to females dying more frequently from their injuries. Seasonal or weather patterns in San Francisco may make this 30-day sample unrepresentative of the overall fatality population. Suffice it to say that this area deserves further study. EDUCATION Information on education is rarely included in the medical record. Our figures, therefore, are limited to our Interview Sample of 32 persons with violence-related injuries. The statistics herein are discrepant with our general clinical impressions of this group. Our figures indicate that 67% have at least a high school diploma. At the same time, we observed that many of our respondents did not comprehend what we considered to be quite straightforward questions or had to have items of the Adjective Checklist defined by the interviewer. Several respondents were unable to obtain jobs because they could not fill out application forms or read instructions. Thirty-six percent had reading problems in school. 50% had failed at least one course, and 32% had failed at least one grade. There are several possible reasons for the discrepancy between grade achieved and general educational level of the interview group. First, there were a number of downwardly mobile 276 respondents in the group. Some were alcoholic and two had become psychotic. These persons were functioning well below whatever educational level they had originally achieved. Several whose fathers had been firemen, electricians, or welders -- often in rural areas -- seemed not to be faring as well themselves in an urban environment. This may have been a function of the complexity of this particular setting, increased competition for jobs, or because of an intervening factor, such as family disintegration, along the way. FIGURE 4 EDUCATION (INTERVIEW SAMPLE) DID NOT COMPLETE HIGH SCHOOL 41% COMPLETED HIGH SCHOOL 56% GRADUATE LEVEL TRAINING 11% MEAN LEVEL OF EDUCATION: 12.19 YEARS EMPLOYMENT While the medical record data on current employment is not considered to be wholly reliable, in this example the figures for the All Injury Admission Group and the Violence-related Interview sample are remarkably similar, if dismaying, in the picture they represent. Weil over half of persons admitted for All Injuries were not employed prior to their injury. Respondents in the Interview Sample were considered to be employed if they were working at all. However, most of those who did have jobs were underemployed, either working many fewer than forty hours per week or unable to get jobs appropriate to their level of skill. One man who was trained as a welder was working as an assistant in a mortuary a few hours a week. FIGURE 5 EMPLOYMENT ALL_INJURY ADMISSIONS INTERVIEW_SAMPLE EMPLOYED 36% 41% UNEMPLOYED 57% 56% UNKNOWN 7% 3% 277 HOUSEHOLD INCOME Income level in our Interview Sample covered a wide range from no income whatsoever to over $30,000 per year. Because interviewees responded in terms of their total household income, these figures may not accurately reflect what they themselves were earning. The design of the questions pertaining to income left room for ambiguous responses. Additionally, the variety of living arrangements was not encompassed by our questions on resources. Four of the 27 persons who responded to this question had no income at all. Fourty-four percent lived on less than $5,000 per year. Sixty-three percent of our respondents earned less than $15,000 per year. A number were living with their families or receiving help from friends. In general, while several persons were indeed holding down full-time, adequately paying positions, most of our respondents were barely "making it" and a few were frankly not. FIGURE 6 HOUSEHOLD INCOME (INTERVIEW SAMPLE) 0-: $2,999 $3 ,000- -$4, 999 $5 ,000 -$9, 999 $10 .000 -$14 ,999 $15 ,000 -$19 ,000 $20 ,000- -$25 ,999 $26 ,000 -$29 .999 30,( )00 OR GREATER UNKNOWN HOUSING (INTERVIEW SAMPLE) Three out of 30 interview respondents had no housing. These persons lived in Golden Gate Park, in cars, or moved about the streets. Several respondents lived at home with parents or alternated between apartments and family homes. A good number lived in apartments with friends or in transient hotels in the seedier sections of the Tenderloin District in San Francisco. Many complained that their living situations were both unsavory and unsafe. These persons attributed the injury experience in part to the conditions of both their housing and neighborhood. 278 25% 13% 6% 9% 9% 3% 3% 16% 16% FIGURE 7 HO (INTERVI HOUSE APARTMENT HOTEL OTHER NONE UK SING W SAMPLE) 25% 41% 16% 6% 9% 3% 279 MARITAL STATUS We have data from the medical record on the current marital status of our group of All Injury Admissions, as well as from our Interview Sample of Violence-Related Injury Admissions. We do not however, have current figures from the General Population of the City and County of San Francisco. The City being what it is, however, any figures might be instantaneously out of date. We do know that there is a large single population in the City, perhaps in large part because of the large subculture of Gay men. Historically, it is also a city of people in transition. Both our population of All Injury Admissions and our Interview Sample of Violence-related Injury Admissions reflect this trend. We would suggest, however, that both samples of the injury population probably reflect more social isolation than the General Population. Fifty-three percent of All Injury Admissions and 59% of the Interview Sample of Violence-related Injury Admissions had never been married. Recall that the mean ages were 36 years and 34 years, respectively. Significant numbers in each group were separated, divorced, or widowed. Several possible explanations of these phenomena occur to us. First, that married persons may be overall a more stable group. Secondly, separation or divorce may lead to depression as well as acting-out of aggressive feelings. Thirdly, volatile persons may, indeed, have higher rates of separation and divorce. In summary, 65% of All Injury Admissions were unmarried and 98% of the Interview Sample of Violence-related Injury Admissions were unmarried. While we would not wish to impose bourgeois social criteria on our study population, we are concerned about the apparent lack of social supports and personal resources among these groups. FIGURE 8 MARITAL STATUS NEVER MARRIED MARRIED SEPARATED DIVORCED WIDOWED UK ALL INJURY 53% 14% 3% 6% 3% 21% ADMISSIONS INTERVIEW 59% 6% 17% 19% 3% 3% SAMPLE OF VIOLENCE- RELATED INJURIES 280 PARENTHOOD AMONG RESPONDENTS (INTERVIEW SAMPLE) Fifteen or 50% of the 30 patients on whom such information was available had children. Average number of children per respondent was 1.06. Two respondents had 3 children and one had 4. Of the 15 parents in our sample, 12 or 80% were unmarried at the time of the interview. Two or 13% were married. Of the currently married parents, 8 or 67% were previously married. Four of these persons were now separated and four were divorced. Four of 27% of the parents in the interview sample had never been married. FIGURE 8A PARENTHOOD AMONG RESPONDENTS (INTERVIEW SAMPLE) CHILDREN NO CHILDREN UNKNOWN 47% 47% 6% MARITAL STATUS OF PARENTS AT TIME OF INTERVIEW MARRIED UNMARRIED UNKNOWN 13% 80% 7% 281 FIGURE 9 MARRIED - SINGLE MARRIED SINGLE ALL INJURY ADMISSIONS 14% 65% INTERVIEW SAMPLE OF VIOLENCE- RELATED INJURY ADMISSIONS 6% 98% B. CHARACTERISTICS OF THE INJURY-PRODUCING INCIDENT TIME OF INJURY PRODUCING INCIDENT We would now like to turn our attention to the circumstances of the injury-producing incident in an effort to outline some of the critical features. With very few exceptions, violence-related injuries occurred during the evening and early morning hours. Information about the details of three injury incidents were unavailable because patients were psychotic or amnesic for the incident, or because the tape recording was faulty. Twelve percent of injury times were unknown. Seventy-nine percent of the 32 injury incidents, however, took place between 6:00 p.m. and 6:00 a.m. Several persons were assaulted by unknown persons for unknown reasons while sitting on doorsteps or leaving friends' houses late in the evening. Most, however, seemed to be a function of what we have come to identify as "Post-Bar Syndrome." FIGURE 10 TIME OF INJURY PRODUCING INCIDENT (INTERVIEW SAMPLE) 0600- 1200- 1800- 2100- 0000- UK 1200 HRS. 1800 HRS. 2100 HRS. 2400 HRS. 0600 HRS. 6% 3% 16% 41% 22% 12% 282 LOCATION OF INJURY PRODUCING INCIDENT (INTERVIEW SAMPLE) In a larger sample of such patients, we would expect to see more domestic violence than we did in this sample of 32 persons. Among women inpatients in particular, abuse among relatives or acquaintances is less often acknowledged. We know from the literature, as well as our clinical consulting experience to the Trauma Service, that these persons are well represented in emergency room populations.<3> Only 12% of our current sample however, received their injury in their own or their antagonist's home. In those that did occur in residences, we could not discern any clear set of events or motives for the stabbing or shooting which followed. Much detail could not be covered in a one-hour interview, nor were many of our respondents accustomed to discussing their behavior, motives, and feelings. We would speculate that several of these situations were drug-related: either the antagonist or patient or both were on drugs, or the content of the dispute was in some fashion related to drug transactions. By far, most of our respondents were injured in or around bars. Sixteen percent were injured inside a bar. Sixteen percent were injured inside another commercial establishment -- a restaurant, dance hall, or the like. Forty-one percent were injured in an outdoor location. While several of the respondents were attacked on the street without apparent provocation, most of the incidents occurred just outside of bars or casinos or comparable drinking establishments. FIGURE 11 LOCATION OF INJURY-PRODUCING INCIDENT (INTERVIEW SAMPLE) Residence 12% Other Indoor Location 3% Bar or Tavern 16% Other Commercial Establishment 19% Outdoor Location 44% Unknown 6% 283 MODE OF INJURY As indicated earlier in our discussion, 35% of All Injury Admissions are a function of interper onal violence: gunshot or stab wounds, or assaults. While we know that some falls, overdoses, and burns are homicidal in nature, information to confirm this clinical impression is difficult to obtain from the medical record. The large number of violence-related injury admissions should be examined with reference to the experiences of other hospitals where the preponderance of injury admissions are due to motor vehicle collisions.<3> Of violence related injury admissions in our survey of All Injury Admissions, 18% result from assaults 13% are from stab wounds, and 4% are the result of gunshot wounds. While gunshot wounds are the fewest in number, they are the most devastating in terms of tissue damage, requiring operative repair in 95% of cases . Of our survey of the 104 Violence-related Injury Admissions in a 30-day period, 46% were the result of assault, 41% were the results of stab wounds, and 12% were gunshot wounds. Of our Interview Sample of 32 patients, 56% suffered assaults, 38% received stab wounds, and 6% had gunshot wounds. Injury Severity Scores were calculated on 30 of the 32 patients. The mean score was 8, indicating severity enough to require an estimated 4 days hospitalization. FIGURE 12 MODE OF INJURY (Violence 35%) Assault Stab Gunshot Auto/Truck Motorcycle Pedestrian Bicycle Fall Overdose Burn Other Unknown 3 Northeast Ohio Trauma Study: Barancik, J., Chatterjee, B. (Traffic 21%) (Other 43%) ALL INJURY ADMISSIONS (*) 18 13 4 9 5 5 2 19 6 3 15 0 VIOLENCE- RELATED INJURY ADMISSIONS (*) 46 41 12 0 0 0 0 0 0 0 0 1 INTERVIEW SAMPLE (*) 56 38 6 0 0 0 0 0 0 0 0 0 I. The Magnitude of the Problem. Greene, Y., Michenzi, E., Fife, B., American Journal of Public Health, Vol. 73, No. 7, 7/83, pp. 746-51 284 NATURE OF INJURY-PRODUCING INCIDENT (INTERVIEW SAMPLE) Identification of the motive for each injury incident in the interview sample was attempted by content analysis of the Unit Act Sequencing section of the Incident Description. In this section, the patient describes the events leading up to and including the incident. He is also asked specific questions about his anatagonist, the presence or absence and behavior of other people, the use of a weapon, verbal interchanges during the incident, and police involvement. Our 32 interview incidents broke down into eight categories. There was occasional overlap among categories. For example, an incident could include both sexual competition and "post-bar syndrome". Where this overlap occurred, we chose what seemed to be the prevailing motive or circumstance. Data were missing in 9% of cases because the patient didn't remember the incident, never knew what hit him, was too brain-dam aged to communicate, or because of faulty tape recording. For 10% of incidents, a motive could not be elicited. In two of these cases, the young men who were injured were stabbed or shot without apparent provocation, and even without words being exchanged. It is possible that they were denying to us or to themselves the nature of the encounter. It is likewise possible from the descriptions of the incidents that the assailants were drug intoxicated. This would acount for the bizarre quality of the encounters. Nineteen percent of the injuries occurred as a function of a robbery. These usually occurred on the street in the vicinity of a bar where the patient had been drinking. Ten percent of injuries were related to drug transactions. These were usually retaliatory in nature. The assailant demanded drugs, money for drugs, or dispatched retribution for a perceived infraction of the guidelines for the transaction. Thirteen percent of the incidents were predominantly racial in flavor. These include Black/White, Hispanic/White, encounters. Nearly invariably these occurred late in the evening after considerable drinking on the part of one or both participants. They generally involved one ethnic group entering what the antagonist perceived as his inviolable territory. The participants were, for the most part, in their early twenties and had had similar previous encounters. Thirteen percent of the incidents seemed to be a function of sexual rivalry or jealousy. Two encounters were accompanied with alcohol, although one was not. Two of the incidents involved males competing for females; in the third incident, a woman assaulted her cousin reportedly to prevent the cousin's involvement with an undesirable man. 285 Ten percent of incidence seemed to be the function of a psychotic person creating a nuisance or involving another in his delusional system. Thirteen percent of incidents were what we have entitled, "Post-bar Syndrome", in which the patient or antagonist or both have been in a bar drinking for a considerable length of time, emerged with inhibitions loosened, often irritable, possessed of an inflated sense of their physical capacities, and provoked or escalated a chance encounter. This category overlaps with several others, but in our estimation, is a significant variable in many of the violent incidents described here. FIGURE 13 NATURE OF INJURY-PRODUCING ENCOUNTER (INTERVIEW SAMPLE) Robbery 19% Drug-Related 10% Racial 13% Sexual competition/jealousy 16% Psychotic: Nuisance/delusion 10% Post-bar Syndrome 13% No Apparent Reason 10% Missing Data 9% RELATIONSHIP OF ASSAILANT In our Interview Sample, 47% of assailants were acquaintances, 50% were strangers, 3% were unknown. This table demonstrates the superiority of the Interview-acquired data over the All Injury Admissions data retrieved from the medical record. For the latter survey, for violence-related injuries, the assailant was not documented in 95-97% of cases. We should emphasize here that we now feel that studies of violence should dispense with such dichotomous categories as Acquaintance versus Stranger and Perpetrator versus Victim. Our findings suggest that these conceptualizations are not valid reflections of the dynamics of these injury incidents. If utilized, acquaintace-stranger and perpetrator-victim data should be examined on continua. The questions are, "In what way and to what degree were the participants "acquainted" or involved with each other?" And, likewise, "To what extent and in what ways did those involved in the incident participate verbally or behaviorally in the injury-producing incident?" 286 FIGURE 14 GSW SW ASSAULT 3/10 3 3 GSW SW ASSAULT 7/10 1 2 GSW 97 SW 95 ASSAULT 95 RELATIONSHIP OF ASSAILANT ALL INJURY INTERVIEW ADMISSIONS SAMPLE ( % ) # % ACQUAINTANCE GSW 3/10 1 47 7 7 STRANi GSW 7/1 1 5 50 10 SELF GSW 5/10 0 0 SW 8/10 ASSAULT 5/10 UNKNOWN GSW 97 0 3 0 1 C. ROLE OF ALCOHOL AND DRUG USE ALCOHOL CONSUMPTION PRIOR TO INJURY (INTERVIEW SAMPLE) Sixty-three percent of 32 members of the Interview Sample acknowledged having used alcohol prior to the injury producing incident. The amount consumed ranged from two beers to "drinking all day." Average number of beers consumed was six. Mean number of glasses of wine was eleven. Average number of drinks of liquor among those who recalled their consumption was six. Of the Interview Sample, 25% denied alcohol use prior to their injury and for 12% no information was available. A comparison of the Interview Sample of Violence-related Injury Admissions to All Injury Admissions with regard to alcohol use prior to injury is possible. Information on All Injury Admissions is collected from the medical record. A patient is considered to be "alcohol-confounded" if there is clinical evidence of intoxication--alcoho1 on breath, altered mental status consistent with alcohol use or laboratory evidence of intoxication, such as positive blood alcohol screen or a serum osmolality above 295 without head injury. 287 Thirty-nine percent of All Injury Admissions were positive for alcohol. Fifty-five percent were negative. Six percent were unknown. The discrepancy between information gathered by interview and medical record review is remarkable in the example. Since toxicology screens are presently done only for purposes of clinical management, many patients who are intoxicated, but for whom screening will not alter treatment, are not tested. An additional explanation for this discrepancy is that the interview does, indeed, retrieve more valid information than does the medical record regarding the patient's behavior. A final contributing factor may be that alcohol consumption is even more highly correlated with violence-related injuries than with the broader range of injuries. FIGURE 15 ALCOHOL CONSUMPTION PRIOR TO INJURY (INTERVIEW SAMPLE) YES NO UNKNOWN 63% 25% 12% FIGURE 15A ALCOHOL INGESTION PRIOR TO INJURY DOCUMENTED IN MEDICAL RECORD (ALL INJURY ADMISSIONS) YES NO UK 39% 55% 6% NON-THERAPEUTIC DRUG USE PRIOR TO INJURY Of our Interview Sample, only 3% of patients acknowledged drug use prior to the injury. One person had smoked 2 marijuana cigarettes. A second person was on methadone which was coded by us as a therapeutic medication. Eighty-one percent of the Interview Sample denied drug use. For 16% pertinent information could not be obtained. These figures are at some variance with those from the All Injury Admissions survey. For that group, 13% had documented drug use. For 87%, there was no documentation in the medical 288 record of drug use prior to injury. Drug use, particularly if it is not intravenous drug use, is probably significantly underrepo r t ed in the medical record because such use poses little immediate risk to the patient who is not already psychotic, and screening will not materially alter the course of acute care. Otherwise, the variance between these two groups is not entirely explicable. Patients who are interviewed may well be more reluctant to acknowledge drug use or involvement in drug transactions than the use of alcohol. Since we know that several of these persons live in an environment where drugs are available, this may well be problematic issue for the interview method. Interviewers were not encouraged to pursue discussion of illegal activities. We do believe that there is more drug intoxication in injury patients than is currently being documented either in the medical record or by interview. As with alcohol use, which we have reason to believe confounds at least 80% of hospital-treated injuries, only a systematic laboratory drug screening of this population will reveal to what extent drug use is associated with violence-related injuries.<4> FIGURE 16 NON-THERAPEUTIC DRUG USE PRIOR TO INJURY INTERVIEW SAMPLE OF ALL INJURY VIOLENCE-RELATED ADMISSIONS INJURY ADMISSIONS 13% YES 3% 80% NO 81% 7% UNKNOWN 16% D. HISTORY OF RECENT STRESS AMONG PATIENTS VIOLENCE-RELATED INJURIES (INTERVIEW SAMPLE) Our Interview Sample was quite extraordinary with regard to the subject's experience with recent stress. The main number of stressors experienced in a 12-month period was 9. Our list of stressors included 56 items relating to job, financial, health, relationship, legal, and housing issues. Only 19% had experienced fewer than 5 stressors during the last 12 months. (This writer reported only 2 stressors during this period). Eighty-two percent had experienced 5 or more stressors. 4 "Voldsulykker og Alkohol," Thaarup, P., Hellenland, H., Kaempe, B., Ugeskrift for Laeger, Volume I2***! Number 34, 8/23/82, pp. 2522-2523. 289 Forty-eight percent had experienced 10 or more stressors. Eleven percent had experienced 15 or more stresors. percent had experienced 15 or more stressors. The most frequently experienced items were unemployment and change of residence (55%). Forty-eight percent had been in jail or lost something of value during the previous 12 months. Thirty-seven percent had been arrested, started a new job, or moved to an improved living situation. Thirty-three percent had experienced legal problems, serious financial difficulties, or illness to a loved one. Thirty percent were laid off from a job, experienced the break-up of a love relationship, or were involved in an argument which led to violence during the previous 12 months. Twenty-six percent of our interview sample had been hospitalized for an illness or injury, had an alcohol problem, or had a relative or friend who was victimized. Twenty-two percent experienced the addition of a relative to their household, the birth of a child in their immediate family, or entered school or a training program. Nineteen percent lost their housing altogether during the last 12 months, were the victim of a violent crime, went seriously into debt, or experienced the death of a family member. Fifteen percent lost a close friend through death, had a serious personal illness or injury, or had an increase of responsibility at work. Eleven percent of the sample had a loved one who underwent a serious operation, or had a family member with an alcohol or drug problem. Seven percent of our sample were fired from a job during the previous year, had to deal with a mentally ill family member, had trouble with in-laws, friends, or neighbors, experienced the return of a child after an absence, had a substantial increase in income, a relative with problems other than health, or a pet which died. Four percent experienced the burglary of their home, trouble at work, separation from a spouse or partner, other marital problems, unwanted pregnancy (self or partner's), hospitalization for mental illness, problems with children, engagement, marital reconciliation, or the completion of school or a training program. Not mentioned as stressors in the last 12 months by any of our respondents were: spouse or partner died, having become widowed, having been divorced, having gotten married, having a child leave home. While our pilot study suffers from having no other group to compare to our interviewers, it does seem to us that the members of our Interview Sample experienced much greater numbers of recent stressors than might members of the General Population. The most prominent issues for our interviewers were unemployment, housing, and experience with the criminal justice 290 system -- arrest, incarceration, as well as other legal problems. Our overall impression of this group was that with few exceptions, most had problems with unemployment or underemployment. Several persons were unable to read well enough to obtain or retain jobs which involved paperwork. Several persons were psychotic or were heavy users of alcohol. This may well have interfered with their ability to work consistently. All but a few had a few saleable work skills. Some of those who were trained in a trade were unable to find work in their industry. Related in great part to the financial duress of our respondents were housing difficulties. The several "street people" were in our Interview Sample were psychotic or severely alcoholic. FIGURE 17 HISTORY OF RECENT STRESS AMONG PATIENTS WITH VIOLENCE-RELATED INJURIES NUMBER OF STRESSORS EXPERIENCED DURING LAST 12 MONTHS (INTERVIEW SAMPLE) MEAN: 9 RANGE: 0-20 FEWER THAN 5 STRESSORS: 19% 5 OR MORE STRESSORS: 82% 10 OR MORE STRESSORS: 48% 15 OR MORE STRESSORS: 11% 291 FIGURE 17A STRESSORS REPORTED BY 27 RESPONDENTS PREVIOUS 12 MONTHS IN ORDER OF FREQUENCY OF RESPONSE # % 15 UNEMPLOYED FOR MORE THAN ONE MONTH 55 15 MOVED TO A NEW RESIDENCE 55 13 IN JAIL 48 13 LOST SOMETHING OF VALUE 48 10 ARRESTED 37 10 STARTED TO WORK ON A NEW JOB 37 10 MOVED TO A BETTER NEIGHBORHOOD 37 9 LEGAL PROBLEM 33 9 SERIOUS FINANCIAL DIFFICULTIES 33 9 SERIOUS ILLNESS TO A LOVED ONE 33 8 LAID OFF FROM JOB 30 8 LOVE RELATIONSHIP BROKE UP 30 8 ARGUMENT WHICH LED TO VIOLENCE 30 7 HOSPITALIZED FOR AN ILLNESS OR INJURY 26 7 ALCOHOL OR DRUG PROBLEM 26 7 RELATIVE OR FRIEND WAS VICTIMIZED 26 6 RELATIVE MOVED INTO YOUR HOUSEHOLD 22 6 ENTERED SCHOOL OR TRAINING PROGRAM 22 6 BIRTH OF CHILD IN YOUR IMMEDIATE FAMILY 22 5 LOST HOUSING 19 5 VICTIM OF A VIOLENT CRIME 19 5 SERIOUS DEBT 19 5 FAMILY MEMBER DIED 19 4 CLOSE FRIEND DIED 15 4 SERIOUS ILLNESS 15 4 SERIOUS INJURY 15 4 GREATLY INCREASED WORKLOAD 15 3 SERIOUS OPERATION TO A LOVED ONE 11 3 FAMILY MEMBER HAD AN ALCOHOL OR DRUG PROBLEM 11 2 FIRED FROM JOB 7 2 FAMILY MEMBER WAS MENTALLY ILL 7 2 PET DIED 7 2 TROUBLE WITH IN-LAWS 7 2 TROUBLE WITH FRIENDS OR NEIGHBORS 7 2 RELATIVE HAD PROBLEM OTHER THAN HEALTH 7 2 INCOME INCREASED SUBSTANTIALLY 7 2 CHILD RETURNED AFTER LONG ABSENCE 7 HOME WAS BURGLARIZED 4 TROUBLE WITH SUPERIORS AT WORK 4 SEPARATED FROM SPOUSE OR PARTNER 4 OTHER MARITAL PROBLEM 4 UNWANTED PREGNANCY 4 HOSPITALIZATION FOR MENTAL ILLNESS 4 MARITAL RECONCILIATION 4 ENGAGEMENT 4 PROBLEM WITH CHILD OTHER THAN HEALTH 4 FINISHED SCHOOL OR TRAINING PROGRAM 4 292 E. RECENT EXPOSURE TO AND EXPERIENCE WITH VIOLENCE A checklist of exposure to and gradations of violence such as: verbal confr shoving, or slapping; use of fist or object; gun during arguments or disputes during the was administered to each respondent in the Encounters with spouse or partner, chil any other people were covered. Differentiation was made between §xp.§lie_iic_e with or participation in these enc Respondents in the Interview Sample were exp_o of 29 incidents of persons screaming or sho during the last year. Respondents Ea£ticip.ated in a mean num of screaming, shouting, or serious arguments d prior to their injury. ®xp_erience with ontations, hitting, and use of knife or previous 12 months Interview Sample. dren, relatives, or ®xp_osure to and ounters. sed to an average uting at each other ber of 16 incidents uring the 12 months FIGURE 18 EXPOSURE TO AND EXPERIENCE WITH VERBAL CONFRONTATION IN LAST YEAR (INTERVIEW SAMPLE) HOW OFTEN HAVE SEEN OTHER PEOPLE SCREAM AND SHOUT AT EACH OTHER? HOW OFTEN HAVE YOU SHOUTED SCREAMED, OR HAD A BAD ARGUMENT WITH YOUR CHILD, PARTNER, RELATIVES, OR OTHER PEOPLE? AVERAGE NUMBER OF INCIDENTS PER PERSON AVERABE NUMBER OF INCIDENTS PER PERSON 29 16 293 EXPOSURE TO VIOLENCE IN LAST YEAR (INTERVIEW SAMPLE) Respondents to the Interview had seen other people hitting each other an average number of 13 times during the last 12 months. Observation of the use of a knife or gun in confrontations between other people occurred for each respondents an average of 3 times during the last year. FIGURE 19 EXPOSURE TO VIOLENCE IN LAST YEAR (INTERVIEW SAMPLE) SEEN OTHER PEOPLE HIT EACH OTHER SEEN OTHER PEOPLE THREATEN TO USE OR ACTUALLY USE A KNIFE .OR_GUN_ON_EACH_OTHER__________ AVERAGE NUMBER OF INCIDENTS PER PERSON: 3 AVERAGE NUMBER OF INCIDENTS PER PERSON: 13 EXPERIENCE WITH VIOLENCE IN THE LAST YEAR (INTERVIEW SAMPLE) Members of our interview sample were the recipients of violent acts an average of 2.8 times during the last year and the initiator of pushing, shoving, slapping, hitting with a fist or object, or confrontation with a gun or knife a mean number of 5.8 times. RECIPIENT FIGURE 20 EXPERIENCE WITH VIOLENCE IN LAST YEAR: AVERAGE NUMBER OF INCIDENTS PER PERSON (INTERVIEW SAMPLE) INITIATE PUSHING, PUSHING, SHOVING FIST OR GUN OR SHOVING, FIST OR GUN OR OR SLAPPING OBJECT KNIFE OR SLAPPING OBJECT KNIFE .8 1.5 .5 3.4 2.3 .1 294 FIGURE 20A EXPOSURE/EXPERIENCE WITH PHYSICAL VIOLENCE IN LAST YEAR (INTERVIEW SAMPLE) AVERAGE NUMBER OF INCIDENTS PER PERSON 22 F. RECENT INVOLVEMENT WITH THE CRIMINAL JUSTICE SYSTEM EXPERIENCE WITH THE CRIMINAL JUSTICE SYSTEM DURING 12 MONTHS PRIOR TO INJURY (INTERVIEW SAMPLE) Forty-one percent of respondents to our interview had been arrested or jailed during the course of the last year. One additional espondent acknowledged being in jail prior to the last 12 months . Forty-four percent of our respondents denied being arrested or in jail during this time. For 15% of our respondents, information regarding recent experience with arrest or jail was unavailable. FIGURE 21 EXPERIENCE WITH THE CRIMINAL JUSTICE SYSTEM (INTERVIEW SAMPLE) YES NO UNKNOWN 41«fc 44% 15% 295 G. HEALTH HISTORY FIGURE 22 RESPONDENTS TREATED FOR PREVIOUS INJURIES (INTERVIEW SAMPLE) Fifty-nine percent of our Interview Sample had been treated in a an emergency room or hospital for previous trauma. Types of injuries included sports injuries, motor vehicle or motorcycle collisions, work-related injuries, and violence-related injuries, such as stab wounds, gunshot wounds, and assaults. Twenty-eight percent of our Interview Sample denied having previous injuries. For 13% of our Sample, no information was available. FIGURE 22 RESPONDENTS TREATED FOR PREVIOUS INJURIES (INTERVIEW SAMPLE) YES NO UNKNOWN 59% 28% 13% 2% RESPONDENTS WITH RECENT HEALTH PROBLEMS (INTERVIEW SAMPLE) Forty-seven percent of our Interview Sample reported recent health problems. These included in order of reported frequency: seizures and cardiovascular problems, respiratory problems; mental illness; hearing problems; and venereal disease. All but two respondents who reported seizures attributed these to overindulgence in alcohol. Most of those with respiratory problems attributed this to smoking. Forty-seven percent of this sample denied recent health problems. For 6% of these persons, there was no information on recent health history available. FIGURE 23 RESPONDENTS WITH RECENT HEALTH PROBLEMS (INTERVIEW SAMPLE) YES 47% NO 47% UNKNOWN 6% FIGURE 23-A HEALTH PROBLEMS IN ORDER OF REPORTED FREQUENCY (INTERVIEW SAMPLE) SEIZURES CARDIOVASCULAR RESPIRATORY MENTAL ILLNESS HEARING PROBLEMS DIABETES SUBSTANCE ABUSE RENAL VENEREAL DISEASE 16% 16% 13% 9% 6% 3% 3% 3% 3% 297 H. PERSONAL AND FAMILY HISTORY RESPONDENTS WITH RECENT HEALTH PROBLEMS (INTERVIEW SAMPLE) Fifty-three percent of respondents experienced the disintegration of their original nuclear family through death, separation, or divorce prior to their own departure from their famiiies. Thirty-eight percent of our respondents grew up in intact families. For 9% of our respondents, there was no information available on family of origin. FIGURE 24 WAS FAMILY INTACT DURING CHILDHOOD? (INTERVIEW SAMPLE) YES NO UNKNOWN 38% 53% 9% SUBSTANCE-ABUSING PARENT OR FAMILY MEMBER? (INTERVIEW SAMPLE) 44% of our respondent's families, there was a substance-abusing parent or family member. Most of these relatives were alcohol abusers. One subject's parent was a polydrug abuser. Two respondents reported siblings with drug abuse problems. Forty-seven percent of our respondents denied the presence of substance abuse in their families. For 9% of our respondents, no information was available on this subject. FIGURE 25 SUBSTANCE-ABUSING PARENT OR FAMILY MEMBER? (INTERVIEW SAMPLE) YES NO UNKNOWN 44% 47% 9% 298 RESPONDENTS' EXPERIENCE OF PRECIPITOUS DEATHS OF FAMILY OR FRIENDS (INTERVIEW SAMPLE) Forty-seven percent of our Interview Sample reported the experience of precipitous deaths of family or friends. Precipitous death is defined for our Pilot Study as death by other than disease or natural causes, that is, death resulting from violence. Of 29 responses to this inquiry, 4 persons reported 3 or more deaths of relatives, siblings, or friends by violence. Forty-four percent of respondents did not report such precipitous losses. For 9% of respondents, this information was unavailable. FIGURE 26 RESPONDENTS' EXPERIENCE OF PRECIPITOUS DEATHS OF FAMILY OR FRIENDS (INTERVIEW SAMPLE) % WHO EXPERIENCED % WHO HAVE NOT % LOSS EXPERIENCED LOSS NO INFORMATION 47% 44% 9% MODE OF DISCIPLINE USED WITH RESPONDENT DURING CHILDHOOD (INTERVIEW SAMPLE) Seventy-two percent of respondents to this question reported the use of primarily corporal punishment for wrongdoing during childhood. Methods included spanking with a hand or hairbrush, whipping with switch or belt, beating, "violent physical abuse," and kneeling on rice. Nineteen percent of respondents received primarily verbal, psychological, restrictive punishments such as grounding, removal of television or other privileges, talking or nagging, and "fear". Two persons reported that no one had disciplined them or that discipline had been delegated to a boarding school. For 9% of our interview sample, Information on methods of discipline were not available. 299 FIGURE 27 MODE OF DISCIPLINE DURING CHILDHOOD (INTERVIEW SAMPLE) PRIMARILY VERBAL PRIMARILY CORPORAL UNKNOWN 19% 72% 9% HOW OFTEN WERE YOU SLAPPED OR HIT BY YOUR PARENTS OR RELATIVES FOR DOING SOMETHING WRONG WHEN YOU WERE GROWING UP? Thirteen percent of our interviewees were punished by physical means "quite often" during childhood. Nineteen percent were slapped or hit "sometimes". Fifty-six percent stated that they were slapped or hit as punishment "only a few times" while growing up. Three percent stated that they were "never" slapped or hit as a chiId. For 9% of our respondents, this information was unavailable. FIGURE 28 HOW OFTEN WERE YOU SLAPPED OR HIT BY YOUR PARENTS OR RELATIVES FOR DOING SOMETHING WRONG WHEN YOU WERE GROWING UP? QUITE OFTEN SOMETIMES ONLY A NEVER UNKNOWN FEW TIMES 13% 19% 56% 3% 9% 300 HOW OFTEN WERE YOU SLAPPED OR HIT BY YOUR PARENTS OR RELATIVES FQR_Ng_GggD_REASON WHILE YOU WERE GROWING UP? Six percent of our sample reported what they perceived as being abused by the adult members of their household during childhood "quite often". Seventeen percent acknowledged being slapped or hit for no good reason "sometimes". Twenty-five percent reported abuse of this sort "only a few times". Forty-three percent denied ever having been "abused" as children. For 9% of the sample, no information was available. FIGURE 29 HOW OFTEN WERE YOU SLAPPED OR HIT BY YOUR PARENTS OR RELATIVES FOR Ng_GOOD_REASON WHILE YOU WERE GROWING UP? (INTERVIEW SAMPLE) QUITE OFTEN SOMETIMES ONLY A FEW TIMES NEVER UNKNOWN 6% 17% 25% 43% 9% HOW OFTEN DID YOUR PARENTS SLAP OR HIT EACH OTHER? (INTERVIEW SAMPLE) Six percent of the interview sample reported that their parents slapped or hit each other "quite often". Thirteen percent of the sample indicated that their parents committed violence against one another "sometimes". Nineteen percent acknowledged their parents' involvement in physical violence "only a few times". physical violence "only a few times". Forty-three percent stated that their parents "never" hit or slapped each other. For 19% of the interview sample, this information was not available. 301 FIGURE 30 HOW OFTEN DID YOUR PARENTS SLAP OR HIT EACH OTHER? (INTERVIEW SAMPLE) QUITE OFTEN SOMETIMES ONLY A FEW NEVER N/A TIMES OR UNKNOWN 6% 13% 19% 43% 19% HOW OFTEN HAVE YOU SLAPPED OR HIT YOUR HUSBAND/WIFE/PARTNER? (INTERVIEW SAMPLE) None of our interviewees reported this kind of encounter on a frequent basis. This is quite understandable since the great preponderance (98%) of our sample were unmarried. Three percent reported slapping or hitting their partner or previous partner "sometimes". Thirty-four percent reported hitting or slapping a partner or previous partner "only a few times". Forty-one percent denied ever slapping or hitting a partner. Several persons had never been in love relationships of any duration and deemed this question "not applicable". These persons along with the few for whom there was no information available on this topic constituted 22% of our sample. FIGURE 31 HOW OFTEN HAVE YOU SLAPPED OR HIT YOUR HUSBAND/WIFE/PARTNER? (INTERVIEW SAMPLE) QUITE OFTEN SOMETIMES ONLY A FEW NEVER N/A TIMES OR UNKNOWN 6% 13% 19% 43% 19% 302 I. DEVELOPMENTAL PROBLEMS AND STRESSORS Our interviewees experienced quite extraordinary numbers of recent stressors. It seemed also that as children they had had to deal with stresses and personal problems with greater frequency and of greater severity than the "average" child. While we have no comparison group to ascertain the relative duress these persons operated under, their own reports of the frequency of problems is included. Most frequently mentioned problem areas were family fragmentation and relationship problems, difficulties in school, financial stress, and social isolation. In the light of our current awareness of the common, but infrequently acknowledged problems of childhood, it would be most desirable in any future studies of this nature to ascertain the relative differences between the problems experienced by this group, the overall injury, population, and a sample of the general population. 303 FIGURE 32 DEVELOPMENTAL PROBLEMS AND STRESORS REPORTED BY 28 RESPONDENTS (LISTED IN ORDER OF FREQUENCY) CHANGES IN FAMILY DUE TO DEATH, SEPARATION DIVORCE, OR MARRIAGE 61% FAILURE IN A COURSE: 50% FEELING ANGRY ENOUGH AT ANOTHER PERSON TO WANT TO HURT THEM: 43% READING PROBLEMS: 36% CHANGE IN FAMILY DUE TO JOB LOSSES, MONEY PROBLEMS: 36% LIVING CONDITIONS, MOVING: 36% FAILURE OF A GRADE: 32% PROBLEMS WITH TEACHERS: 32% FEELING DEPENDENT: 32% FEELING HELPLESS: 29% PROBLEMS WITH PEOPLE MAKING FUN OF YOU: 25% FEELING THAT YOU DIDN'T WANT TO GO ON LIVING: 25% FEELING LONELY: 25% RUNNING AWAY FROM HOME: 21% ISOLATION FROM COMMUNITY BECAUSE OF CULTURAL CHARACTERISTICS: 18% FEELING HOPELESS: 14% FEELING POWERLESS: 14% PROBLEMS MAKING FRIENDS: 14% PROBLEMS KEEPING FRIENDS: 14% SEXUAL ABUSE BY ANOTHER PERSON: 7% ISOLATION FROM COMMUNITY BECAUSE OF RELIGIOUS BELIEFS: 7% ISOLATION FROM FAMILY AND COMMUNITY BECAUSE OF SEXUAL ORIENTATION: 7% FEELING YOU WOULD NOT BE ABLE TO GET WHAT YOU WANT OUT OF LIFE: 7% PHYSICAL ABUSE BY ANOTHER PERSON: 4% 304 J. COST OF INJURIES Anyone who remains unimpressed by the destructive impact injury has on the lives of those injured and their families will not fail to react to the cost of violence in dollar terms. In estimating monetary costs, one must include the hospital charges and physicians' fees for the initial hospitalization and any subsequent hospital admissions for reconstructive surgery or for complications. The cost of outpatient visits, equipment and medications must be considered. Visiting nurses and rehabilitative services are often necessary. These costs now amount to over 7% of our national health care expenditures. Other financial costs are in property damage, insurance payments, and in foregone earnings. The latter equal 2.3% of our Gross National Product.<5> The cost of the initial inpatient stays of nine persons with gunshot or stab wounds, or assaults are here listed. All are patients included in our interview sample. These costs include hospital charges and physicians fees only. They do not cover the cost of readmission for further reconstruction or for complications. These figures, likewise, do not reflect the cost of follow-up clinic visits, equipment, or outpatient medications. Our sample of 9 patients with gunshot, stab wounds or assaults was not randomly selected. Since all hospitalizations occurred within the last six weeks, data on many persons, particularly those admitted toward the end of the study period, had not reached medical records and billing terminals. However, since our examples were chosen solely because of the availability of pertinent data, we have no reason to feel that they are not representative of the group. Three cases of assault were included. Mean length of hospitalization for these injuries is 4.3 days. The average Injury Severity Score is 13. (The range includes a low score of 0 to a high of 75). Mean cost of initial hospitalization is $7762. Total cost of the care of these three cases for such assaults was $23,285, an average per patient per day of $1791. Three cases of stabbing are included. Mean length of initial hospitalization for this group is 3.3 days. The average Injury Severity Score is 24. Mean cost of initial hospital care of these persons was $6402. The total cost of hospitalization for three cases of stab wounds is $19,215 or an average per patient per diem cost of $3,040. "Economic Costs of Trauma, United States, 198^," Munoz, E., Journal of Trauma, Volume 24, Number 3, 1984. 305 Three cases of gunshot wounds are included. Mean length of hospitalization is 6 days. Mean Injury Severity Score is 2. Mean cost of hospitalization is $4853. Total cost for three patients is $14,561 or an average per patient daily cost of $909 . The cost of care for one month's worth of violence-related injuries was $655,595. If this month is representative, the yearly bill for the acute care of wounds suffered through violence is almost eight million dollars. Can we afford to deny this as a problem? FIGURE 33 COST OF INITIAL INPATIENT HOSPITALIZATION FOR PERSONS WITH INJURIES RESULTING FROM GUNSHOT, STAB WOUNDS, AND ASSAULTS NUMBER OF INJURY DAILY TOTAL COST OF HOSPITAL DAYS SEVERITY SCORE COST HOSPITAL CARE ASSAULT-1 4 4 $851 $3404 ASSAULT-2 5 4 $2080 $10399 ASSAULT-3 4 30 $2371 $9482 TOTALS 13 38 $5302 $23286 MEANS 4.3 13 $1767 $7762 STABBING-1 1 21 $6240 $6240 STABBING-2 5 9 $1449 $7243 STABBING-3 4 41 $1431 $5724 TOTALS 71 $9120 $19207 MEANS 3.3 24 $3040 $6402 GSW-1 7 1 $ 921 $6448 GSW-2 7 1 $ 675 $4724 GSW-3 3 4 $1130 $3389 TOTAL 17 6 $2726 $14561 MEANS 6 2 $ 909 $4853 HOSPITAL CHARGES AND PHYSICIANS* FEES, 105 ADMISSIONS IN 30 DAYS: $665,595. YEARLY COST OF HOSPITAL CHARGES AND PHYSICIANS' FEES AT S.F.G.H FOR GUNSHOT WOUNDS, STAB WOUNDS, AND ASSAULTS: $7,987,140. VII. PROFILE OF CHARACTERISTIC VIOLENCE-RELATED INJURY PATIENT A profile of the characteristic patient in our pilot study of violence-related injuries is a thirty-five-year-old, single, Black, male. He has a high school diploma, which he achieved with some difficulty. He has few marketable vocational skills. He lives in a house or apartment with family, extended family, or friends and has moved at least once during the last year. His household income is less than $15,000. per year. Our patient is unemployed, underemployed, or not presently 306 working in the industry for which he trained. This individual abuses alcohol in some fashion by drinking too much on weekends, binge drinking, or chronic daily use. He has experienced an exceptional number of stressors during the twelve months prior to his injury. Predominant areas of stress are employment, housing, incarceration, and the loss of valued property. The representative patient in our sample experienced the fragmentation of his family due to death, separation, or divorce while growing up. He has lost at least one relative or friend by violent death. As a child, he was affected by the loss of a parent, problems in school, and recalls feeling angry enough at someone to want to hurt them. Our patient has been treated for previous injury and currently experiences health problems such as seizures, cardiovascular, or respiratory problems. Principal mode of discipline during childhood was spanking or whipping. He recalls being hit "for wrongdoing" and "for no good reason" only a few times. His parents used violence with each other only a few times, as has our patient when he has been in a love relationship. Our patient has observed verbal confrontation or violence approximately five times per month during the last year. He has had experience as the recipient or initiator of violence slightly less than once a month during the last year. Personal adjustment and self-control scores on Cough's Adjective Checklist are slightly below average, while measures of aggression and self-esteem are slightly above the mean for this sample of 27 respondents. VIII. DISCUSSION Neither our interview sample of 32 patients nor our survey of 105 Violence-related Injury Admissions in this thirty-day period, can be generalized to other facilities or to other segments of the violence-re la ted injury population without caution. While San Francisco General Hospital receives most of the trauma cases occurring in the County, we still know relatively little about those persons treated by other facilities and those treated, but not admitted, by our Emergency Service. Likewise, as this was a study of the feasibility of utilizing interview methodology, casefinding for our interview sample was as systematic as was possible, but not random. A definitive study of this population should entail large samples covering daily and seasonal variations. There should certainly be one or more comparison groups such as a sample of All Injury Admissions, injury non-admissions, or the General Population. Thirdly, some of the accompaniments of violence-related injuries may reflect effect-effect relationships, that is, result, as did the injury, out of a more complex and subtle set of determinants. Our pilot study is descriptive and any inferences regarding causal relationships should be made cautiously. Our recommendations reflect this conservatism. We 307 look forward to the opportunity to execute a more powerful study of these phenomena. This initial interview format was broad. We can now see clearly certain areas in which more detail would have been desirable. Still, we were not sure at the outset who we were interviewing and did not want to be destructively invasive with a vulnerable population. Given the neediness of this patient group, monetary compensation might have increased cooperation with the interview protocol and should be a part of any future study. And, while our general perceptions were that our patients were reliable informants, there is perhaps a good deal which they did not tell us for one reason or another. For example, little detail was gathered about drug use which clinical staff perceive to be a major issue in violent injuries. Pilot testing of pertinent questions would be helpful, as would routine toxicology screening for drugs on a random sample of violence-related injury admissions. As interviewing proceeded, conceived distinctions between acquaintance and stranger, as well as victim and perpetrator began to break down. We now perceive these categorizations as oversimplifications which obscure the real dynamics of the incident. Future studies should focus on violence as the principal feature and seek to elicit to what extent the participants were aware of or know each other and in what manner personal characteristics or those of the patient's lifestyle or living circumstances contributed to the injury-producing incident. Finally, we need more information about the participants who were not hospitalized. IX. RECOMMENDATIONS Our findings illustrate that in addition to being a major public health problem, injury, particularly violence-related injury, is a serious mental health problem. The high proportion of injury repeaters, as well as the high incidence of suicide, are outstanding manifestations. Essential to the formulation of effective intervention and prevention programs is a clearer understanding of the problem. Systematic studies of the dynamics of violence-related injuries are critical. Particularly important would be studies of the incidence and role of alcohol in violence-related injuries as well as the incidence and role of drug use and drug transactions in such injuries. The basic survival, as well as functional, needs of this group of patients must be better appreciated, as must impediments inherent in the individual, his situation, and in the society at large, to the meeting of these needs. As in the areas of neoplastic and cardiovascular diseases, the roles of personality factors which dispose toward injury should be explored. At the hospital and emergency service levels, staff must be trained to identify and document pertinent data regarding 308 violence-related injuries of all types. As in the review of other disease entities, an historical and dynamic approach is required. Protocols for the identification of violence-related injuries are being developed in medical centers. These should be refined and adapted for use in various settings, and introduced by qualified staff into facilities receiving significant numbers of these patients. Treatment of high-risk adults must begin in the emergency room and hospital inpatient settings. Persons with violence-related injuries are not frequently connected in a meaningful way to other kinds of agencies. Those to which they do turn, such as General Assistance programs, are unable to intervene with the broad range of these persons' needs. Experimental identification and treatment programs should be established for emergency room and inpatient use. Since these patients do not present with formal psychiatric diagnoses, programs should include more subtle casefinding media than psychiatric referral systems. Well-trained psychiatric or clinical social workers, utilizing routine screening methods with all trauma patients, are ideally suited to this task. Treatment modalities should be holistic, encompassing the range of social, vocational, legal, medical, and psychologic needs that these persons present. Focus of counselling efforts should be the strengthening of problem-so lv ing skills, increasing of the sense of efficacy in dealing with the world, the capacity to reflect on the consequences of one's behavior, and the exploration of alternatives to violent methods. The sources of such "acting-out" behaviors--rage and depression-- must be explored within the confines of a nurturing but structured relationship. This latter aspect requires culturally sensitive clinicians, the capacity for outreach into the community, and the ability to follow these patients over time, and low cost or sliding scale for payment systems. Structural interventions must include: the Identification of learning problems in children otherwise at high-risk for antisocial behavior. Many of our respondents expressed a sense of helplessness and resignation regarding the possibility of their ever obtaining meaningful work skills. Casefinding in schools with the use of such protocols might allow for earlier intervention with youngsters who are at risk. School counselling programs have been largely barebones or discontinued during the last two decades. This policy might be re-evaluated. Literacy programs should target this population and focus on applied skills. Viable job training programs need to be established. While the secondary school education system is generally under fire, the most serious omissions in our opinion, have been made in the cases of children who may not wish a liberal arts education beyond high school. Likewise, the elimination of such programs as the Job Corps, which certainly had its problems, means that there is no avenue in our society at present, save the military, for the structuring of the growth for our late adolescent and young adult populations. Employment opportunities need to be increased at the 309 semi-skilled and unskilled levels. The conduits to employment should be integrated and more visible to this group. Vocational assessment and training should be likewise available and marketed positively to youth. It seems awkward still to be repeating that discrimination against minority groups in hiring must be faced and dealt with more definitively. Societal changes need to be outlined and energized. Socialization in violence within the family through corporal punishment, frank child abuse, the death penalty and particularly in the media need to be challenged. Violence is now perceived not only as a coping mechanism, especially among males, but as a legitimate means of achieving one's objectives. An ongoing culture of compulsive masculinity with its accompanying denigration of women, Blacks, Gays, and others who are not depicted in media commercials for Camel cigarettes supports this. Positive images of Black and other minority males are critical. The women's movement has succeeded in changing role models for many American women, and we believe that similar improvements can occur for men. This must include the elimination of role models which communicate that violence or violently-oriented activities, such as boxing and the violent aspects of hockey and crime are the only routes to success. Effectiveness in other kinds of activities such as those of Edward Brooke, Willie Brown, Bryant Gumble, Ed Bradley, the functioning construction worker, military officer, artist, computer programmer, business executive, and shopkeeper must be portrayed as desirable and possible. Advertising and the media must be used constructively. Finally, failing fami1ies--Black, White, Hispanic, Native Ame r i c an --need to be shored up and supported in ongoing ways. Responsible day care is crucial. Forms of grants in aid which are both more supportive than the current programs, and at the same time encourage the incentive to achieve, perhaps on the inverted income tax model, will be an essential part of the process. The problem of teenage pregnancy, which is a symptom of the general hopelessness of American minority youth, must be taken seriously and understood empathetica1ly. Programs aimed at enhancing the lives and skills of teenage parents, school dropouts, immigrants, and children and teenagers with behavior or other special problems must be reformulated and reinstated. Structured nurturing must be made available to all children and the burden on the remaining parent in the one-parent family relieved. Integrated programs for children and functioning elderly might be entertained. Alcohol and drug treatment programs must be improved and expanded. It is our clinical experience that many persons who wish to discontinue their substance abuse endure unbearable waits for admission to programs or are found unacceptable for admission to programs which do not acknowledge or provide for the widespread existence of polydrug abuse. Housing is identified by our Interview Sample as a major stressor. We are aware, however, that homelessness is an 310 underacknowledged social problem in the country as a whole and particularly in urban areas. Decent and aesthetically satisfactory low-income housing, low-income hotels, and residential centers for those who cannot live wholly independently must be made available. The frustration and emotional wear and tear of unsafe and inadequate living situations on vulnerable segments of our population could be significantly reduced by earnest attention to their special housing needs . Widespread fragmentation of the nuclear family, the dissipation of extended family systems, and the deterioration of our schools has resulted in a vacuum insofar as the inculcation of basic social values is concerned. Attention must be given to how such concepts as social responsibility, moral judgment, and concern for one's community are to be conveyed to our youth. Likewise, among adults--particularly those involved with youth--there are similar gaps in the skills of compromise, negotiation, as well as other problem-solving skills. We recommend such approaches as conflict resolution classes to law enforcement personnel and the increase in interpersonal skills among teachers, public agency representatives, and others who come in contact with youth and vulnerable populations. It has been demonstrated that the experience of hope and a sense of control over one's environment correlates with decrease in aggressivity.<6> The creation of organizations or mechanisms which would provide youth with the opportunity to effect social change in areas important to their own lives would both reduce the burden on government and incorporate a large portion of our population into the mainstream. The list is long and the problems are complex. The alternative to confrontation with these apparently elusive social problems, as we have seen, is the cost of the lives and livelihoods of our youth. Whether in dollars or in tears or both, we pay. "Ideology and Injury Prevention." Whitman, S. Paper presented at the Workshop on Assaultive Violence and Homicide: New Directions for Hospital-based Research and Prevention, 12/84, Philadelphia. 311 o U. S. GOVERNMENT PRINTING OFFICE : 1986 491-313/44710 NATIONAL LIBRARY OF MEDICINE NLfl QOb?D45b 1 NLM006704569