Community Risk Factors 1 Community Risk Factors for Hate Crimes: Race/Ethnic and Economic Change

Research on the relationship between community race/ethnic and economic change and the base rates of hate crimes has been rarely studied in the social sciences. The present study examined the role of race/ethnic and economic change in Los Angeles between 1990 and 2000 to determine their relationship to hate crime occurrence. Data collected from Los Angeles hate crime reports, including victim and offender race/ethnicity, the level of severity, and the level of bias, were combined with census data for the 1990 and 2000 censuses for race/ethnic and economic change in the corresponding census tract in which the hate crime/incident occurred. No relationship was found between economic change and hate crimes. While differences among victim race/ethnicity (White, African American, and Hispanic) and their corresponding race/ethnic change (decreasing, stable, or increasing) were largely not significant, there were significant differences between African American and White offenders and their corresponding change in race/ethnic population. Community Risk Factors 4 In 1986, a group of 20 White teenagers brutally attacked three African American men in Howard Beach, New York (Levin & McDevitt, 2002). The teens were motivated by the belief that the African American men did not belong in their neighborhood and needed to be driven out. Hate crimes are often shocking acts committed by a prejudiced few to rid their community of undesirable groups. This case called attention to the severity of hate crimes, as have many incidents since, such as the brutal murders of Matthew Shepard and James Byrd. These extreme cases are but a few of the many hate crimes that frequently occur in our present time. While many hate crimes are less severe in physical damage than murder, all hate crimes and hate incidents have the potential to inflict great emotional harm not only on the victims, but on the whole community of the victim as well. Because of the serious and far reaching effects of hate crimes, many researchers have examined both the precipitating factors and the consequences of hate crimes in an effort to understand and attempt to prevent hate crimes. Although crimes committed with bias intent are hardly a recent phenomenon, legislation attempting to combat hate crimes is relatively new. In 1990, congress passed the Hate Crime Statistics Act (HCSA), which required the federal government to collect data on hate crime incidence in the United States (McVeigh, Welch, & Bjarnason, 2003). The HCSA classifies hate crimes as “crimes that manifest evidence of prejudice based on race, religion, sexual orientation, or ethnicity, including where appropriate the crimes of murder, non-negligent manslaughter, forcible rape, aggravated assault, simple assault, intimidation, arson and destruction, damage, or vandalism of property.” The HCSA was amended in 1994 to include crimes committed with bias against individuals who are mentally or physically disabled. Community Risk Factors 5 Several problems plague the uniform collection of data on hate crimes in the United States. One problem of coming to an accurate gauge of hate crime incidence in the United States is the varying definitions of hate crimes across states. Not all states have laws against hate crimes and among those that do, some criminalize hate acts themselves, while other states only add an enhancement to other crimes which include a hate element. Additionally, some of the states with laws against hate crimes do not include bias against sexual orientation, suggesting that an acceptance of bias against homosexuals is prevalent in these states (Johnson & Byers, 2003). Accurate methods of collecting and labeling hate crimes also vary greatly between states. While California reported 1472 hate crimes in the year 2003, other states, such as Alabama and Mississippi only reported one hate crime each (U. S. Department of Justice, 2004). Clearly these values do not represent the actual prevalence of hate crimes in these states. One problem is that the Hate Crime Statistics Act does not require individual law enforcement agencies to collect data on hate crimes. Alabama and Mississippi each have only one police department reporting to the federal government. Another problem in compiling an accurate account of the frequency of hate crimes is the reluctance of victims of hate crimes to report the occurrence to the police. Herek, Gillis, and Cogan (1999), studied lesbian, gay and bisexual adults’ experiences with victimization by surveying 2,259 individuals. Their study found that only 35-46% of the subjects who were victims of hate crime reported the incident to the police, while 61-72% of the subjects who were victims of other crimes reported the incident to the police. Victims of hate crimes may be fearful of the repercussions of making their victimization Community Risk Factors 6 known. Despite these shortcomings, statistics on hate crimes reported to law enforcement can provide valuable information to researchers. The disparity of hate crime criminalization among states and their law enforcement agencies exemplifies the importance of generating research that clearly identifies the unique qualities of hate crimes and the need for unified labeling, prosecution, and prevention of hate crimes. Some critics have questioned legislation against hate crimes. Jeff Jacoby asserts that hate crime laws punish opinions, not actual crimes, and therefore deny a basic human right to freedom of thought (Jacoby, 2002). However, hate crime legislation does not punish the thoughts of the offender, but rather the greater amount of harm caused by the bias act. Nolan and Akiyama (cited in Nolan, Akiyama, & Berhanu, 2002) cite several benefits of hate crime legislations: (HCSA) would... (a) raise the public’s awareness about the nature and extent of hate crimes; (b) provide a baseline for research and program development; (c) help support the development of effective hate crime legislation; (d) provide law enforcement with information to help them become more effective in working with communities to combat hate crime; and (e) encourage victims to come forward and get the support and assistance they may need. (p. 137) The importance of hate crime legislation has been supported by numerous studies. Research has found that hate crimes are often more violent and instrumental in nature than other crimes (Dunbar, 2003). Hate crimes are also more psychologically damaging to victims than other crimes over a longer period of time (Herek, Gillis, & Cogan, 1999; McDevitt, Balboni, Garcia, & Gu, 2001). Community Risk Factors 7 Edward Dunbar (2003) examined the distinct characteristics of hate crime offenders in order to aid the prosecution of these offenders, as well as to contribute to our knowledge of hate crime offenders as a distinct class of criminals. In this study, Dunbar reviewed the criminal records of 58 hate crime offenders convicted in Los Angeles County between 1995 and 1997. These offenders were coded on several scales that measured the offender’s bias, type of aggression, crime severity, risk factors for criminal activity, and psychopathology. Analysis of the data revealed variation in level of bias among offenders. Of the offenders studied, those high in bias motivation only committed race/ethnicity biased crimes. In addition, the study found a majority of the hate crime offenders committing instrumental aggression, as measured by the Cornell Aggression Index (Cornell, et al. 1996). The analysis also revealed that victims of hate crime offenders are more likely to have had no prior relationship with the offender. Dunbar’s study found that hate crimes are more likely to involve multiple perpetrators rather than individual offenders. This study additionally linked hate crime offenders to substance abuse, poverty, and a dependence on violence. These findings call attention to the distinctiveness of hate crime offenders as a specific class of criminals who can be particularly harmful to the victims of their hatred. Dunbar’s study supports legislation for hate crime enhancements. In an effort to categorize the motivations of hate crime offenders, McDevitt, Levin, and Bennett (2002) developed a typology of hate crime offenders consisting of four specific motivations believed to trigger hate crime offenses. The authors reviewed 169 hate crimes with know offenders reported in Boston, Massachusetts. From these cases, McDevitt et al. identified common motivations for committing hate crimes. The Community Risk Factors 8 motivations described are thrill, defensive, mission, and retaliatory. The authors found that hate crimes motivated by thrill-seeking are usually committed by youths with a minor and often fleeting commitment to their bias beliefs. Defensive hate crimes are committed when the perpetrator believes that he/she must defend their neighborhood from a perceived threat created by the out-group targeted. Committing offenses because of a desire to rid the world of a specific population is described by the authors as a mission motive for hate crimes. Retaliatory hate crimes are performed in order to obtain revenge for a real or perceived incident aimed at a member of the community of the offender. This type of offender is responding to a specific event, and not solely reacting to the presence of a disliked group. Although it is questionable how well actual offenders fall into these specific categories, this typology can help in determining the relative risk of the offenders for recidivism and the corresponding sentencing necessary. Given the distinct characteristics of hate crime offenders, it is clear that hate crimes are a unique type of crime in need of special considerations in regards to the punishment of hate crime offenders. The difference between bias and non-bias crime becomes especially clear when considering the damage inflicted on the victims of hate crimes. Though few studies have looked at the effect of hate crimes on the victims, initial studies suggest that victims of hate crimes often suffer greatly from their victimization. Herek, Gillis, and Cogan (1999) have examined the emotional damage to victims that results in the aftermath of hate crimes. In Herek et al.’s survey of lesbians, gays, and bisexuals mentioned above, they questioned victims of both hate crimes and non-bias crimes about their victimization experiences, psychological well-being, worldview, and victimization-related beliefs. The study revealed a greater level of emotional distress as Community Risk Factors 9 well as a more negative world view among gay and lesbian hate crime victims in comparison to non-bias crime victims. These emotional problems continued to affect the gay and lesbian victims of hate crimes for a longer period of timefor some as long as five yearsin comparison to victims of non-bias hate crimes who on average needed about half as much time to recover from the victimization. These differences were not found among the bisexual victims of hate and non-bias crimes. Although the study did not examine victims of crimes due to other biases such as race/ethnicity and religion, the study highlights the need for specialized psychological treatment for victims of hate crimes and the important distinction between hate crimes and other non-bias crimes. Another study comparing the differences between bias and non-bias crime victims was conducted by McDevitt, Balboni, Garcia, and Gu (2001). This study looked specifically at victims of bias and non-bias related violent assault by sending a survey to victims of these crimes located in Boston, Massachusetts. The survey questioned the victims about the crime itself, the victim’s demographic information, the psychological effects of the assault, the response of their families and the community to the crime, and their feelings about the police and prosecutors’ handling of their case. Although the researchers received a low response rate (about 23% of the bias assault victims and 11% of the non-bias assault victims), the data provides tentative support for the conclusions of the Herek et al. (1999) study. The victims of bias assault were found to have a significantly greater amount of nervousness, depression, difficulty with concentration, suicidal feelings, and difficulty in avoiding thoughts about the victimization in comparison to the non-biased assault victims. Both of the Herek et al. and the McDevitt Community Risk Factors 10 et al. studies emphasize the importance of hate crime legislation due to the harmful and enduring nature of hate crime victimization. One of the key distinctions between hate crimes and non-bias crimes is the secondary effects of the hate crime (American Psychological Association, 1998). A hate crime not only targets the individual victim, but also is meant to convey a message of threat to the entire community the victim represents. McDevitt refers to the community as the “secondary victims” of the hate crime. The messages of hate, as stated by Levin and McDevitt (2002) “are in their intended effect very much like acts of terrorism, meant to send a signal by means of fear and horror.” Additionally, these effects differ from the effects of non-bias crimes in that it is difficult for hate crime victims to change inherent characteristics such as the color of their skin in order to avoid future victimizations. In contrast, victims of non-bias crimes can decide to change the behavioral patterns that might have led to their victimization. Susan Fiske (2002) reviewed the research to date on bias and intergroup conflict to develop a cohesive picture of the nature of biased individuals. The paper presents two types of biased individuals: those with bias that is subtle and those with extreme bias. Fiske stated that most bias is the subtle type which is automatic and often present in “well-intentioned moderates.” Extreme bias, while less common, often results in aggressive behavior. In describing the apparent causes of extreme biases, the author concluded that overall, research points to economics, however: The state of people’s own wallets does not motivate their degree of prejudice. Community Risk Factors 11 Instead, the most reliable indicator is perceived threat to one’s in-group. Group threat (e.g., high local unemployment) correlates with extreme biases against outgroups perceived to be responsible. (p. 127) The connection between economics and the incidence of hate crimes is still debated in current research (Green, Glaser, & Rich, 1998; Hepworth & West, 1988). As the quote above suggests, studies of economic conditions and hate crimes have focused on the community in which the hate crimes occur and instead of the poverty of individual offenders. In addition to studying the unique characteristics of hate crime offenders and victims, researchers have tried to identify some of the factors that put a community at risk for hate crime occurrences. Two common risk factors proposed are economic competition with its related frustrations, and changes in ethnic/racial populations. The assertion that economic difficulty can lead to hate crimes is based on the frustration-aggression hypothesis first proposed by Dollard, Miller, Doob, Mowrer, and Sears (1939). The frustration-aggression hypothesis claims that “the occurrence of aggressive behavior always presupposes the existence of frustration and, contrariwise, that the existence of frustration always leads to some form of aggression” (Dollard et al., 1939). The authors suggested that economic conditions, such as poverty, could sufficiently cause frustration resulting in aggressive acts. To explain the disparity between the number of frustrations we face each day and the lack of aggression exhibited by most individuals, the authors stated that aggressive reactions “may be temporarily compressed, delayed, disguised, displaced, or otherwise deflected from their immediate and logical goal” Dollard et al., Community Risk Factors 12 1939). The displacement of economic frustrations to race/ethnic minorities may be one explanation for the occurrence of hate crimes. Hovland and Sears (1940) tested the frustration-aggression hypothesis by comparing the incidence of lynching to the value of cotton. The researchers hypothesized that when economic conditions worsened, this frustration would lead to the commission of aggressive acts, specifically White lynching African Americans. As the measure of aggression, base rates of African American lynchings occurring in 14 southern states between 1882 and 1930 were collected from The Negro Yearbook. Because the economy of the south was heavily dependent on cotton, the authors concluded that changes in the value of cotton are an appropriate measure of economic conditions during this time period. They collected both the farm value of cotton and the per-acre value of cotton in the 14 states during 1882 to 1930. Hovland and Sears also measured the economic conditions with the Ayres index, a measure which “includes weighted individual measure of consumption, production, construction, imports, exports, and prices” (Hovland & Sears, 1940). Hovland and Sears (1940) performed tetrachoric correlations on the state by state comparisons of lynching base rates and the cotton values culled from these states, as well as between the total number of lynchings in the United States and the Ayres’ index between 1882 and 1930. The researchers’ analysis revealed a high correlation between decreases in the value of cotton and the incidence of lynchings on both the state and national level. The authors concluded that those frustrated by their economic conditions could not commit aggressive acts against symbolically appropriate representations of their economic troubles, such as the wealthy, because of the legal repercussions of such Community Risk Factors 13 crimes. The aggression was instead displaced to the African Americans, who were considered easy targets in a climate that was for the most part accepting of bias crimes against African Americans lynchings. Several attempts have since been made to reanalyze and extend the Hovland and Sears (1940) study (Hepworth & West, 1988; Tolnay, Deane, & Beck, 1996; Green, Glaser, & Rich, 1998). The development of more sophisticated correlational methods than those available in the 1940s allows researchers to investigate the Hovland and Sears data with greater accuracy. The prevalence of references to the original study and its support of the frustration-aggression hypothesis in social psychological literature also encourages the re-examination of its premises. Hepworth and West (1988) replicated the Hovland and Sears (1940) study comparing lynchings to economic conditions. The authors used the same measure of aggression-the number of African American lynchings occurring between 1882 and 1930. Frustrations were measured through the Ayres index, the farm value of cotton, and the per-acre value of cotton, just as in the original study. Hepworth and West decided to also look at the relationship between White lynchings (calculated by taking the total number of lynchings in the years 1882 to 1930 and subtracting the number of African American lynchings in this time period) and the above measures of economic conditions. Hepworth and West first replicated the methods used by Hovland and Sears to ensure that the original analysis wasn’t faulty. The authors also conducted contemporary time-series analysis of the original data. They proposed that a more accurate measure of correlation between lynchings and the economic indices could be obtained with modern methods of analysis. The time-series analysis used by Hepworth and West controlled for variables Community Risk Factors 14 the researchers believed might have skewed the results of the original study including trend, seasonality, and serial dependency. The replication performed by Hepworth and West (1988) revealed similar correlations to those found in the original Hovland and Sears (1940) study for the relationship between African American lynchings and both the Ayres index and the farm value of cotton. However, they were not able to replicate the correlation between the peracre value of cotton and African American lynchings. The authors believed that this may be the result of an error in calculation in the original study by Hovland and Sears, and further concluded that for the most part, the Hovland and Sears data had been accurately replicated. By performing the contemporary time-series analysis, Hepworth and West found smaller correlations between the three economic measures and African American lynchings than those found by Hovland and Sears. The authors concluded that Hovland and Sears had overestimated the relationship between economics and lynchings; however, they still believe a significant relationship exists. The time-series analysis of the relationship between White lynchings and economic conditions revealed a negative correlation, adding further support to the frustration-aggression hypothesis. Overall, Hepworth and West concluded “... the present re-analysis together with other laboratory and naturalistic data ... provide support for the displacement of aggression to minority group members under difficult economic and other stressful conditions.” A more critical re-analysis of the Hovland and Sears (1940) study was performed by Green, Glaser, and Rich (1998). Green, Glaser, and Rich replicated the Hepworth and West (1988) time-series analysis performed on the original Hovland and Sears data. The researchers also used the Hovland and Sears’s data on total African American lynchings Community Risk Factors 15 in 14 southern states and the economic indices of the Ayres index, the farm value of cotton, and the per-acre value of cotton. In addition, Green, Glaser, and Rich extended the time-series analysis of lynchings and economic indices beyond the Great Depression, from 1882 through 1938. By extending the comparison of lynchings and economic conditions through the Great Depression, and using other statistical methods of analysis, Green, Glaser, and Rich (1998) found very different results than previous studies (Hovland & Sears, 1940; Hepworth & West, 1988). Green, Glaser, and Rich found that only when they used the economic measure of the Ayres index were they able to find a negative relationship with the incidence of African American lynchings. Their analysis also found no relationship between economic conditions and all but one of several measures of the type of lynching. Because most measures of the relationship between lynchings and economic deficits failed to provide a significant relationship, Green, Glaser, and Rich question the assertion that the frustration-aggression hypothesis is supported by research on economic conditions and intergroup violence. To further test the hypothesis, they conducted a contemporary study of economics and hate crimes. Green, Glaser, and Rich (1998) performed a present day study looking at economic conditions and hate crime incidence in New York City between 1987 and 1995 to provide further critical analysis of the frustration-aggression hypothesis. The study of hate crime incidence and economic indicators was conducted in four New York City boroughs: Brooklyn, Queens, Manhattan, and the Bronx. The hate crime data consisted of incidents reported to the New York City Police Department’s Bias Incident Investigative Unit (BIIU) between 1987 and 1995. Economic conditions were measured using the Community Risk Factors 16 overall monthly unemployment rate for the four boroughs. The researchers looked at the overall unemployment rate rather than a more specific division because the rates among the four boroughs were very similar. Green, Glaser, and Rich were interested in finding a temporal connection between economic connections and hate crime incidence. Analysis of the data revealed no correlation between the unemployment rate and hate crime incidence. Only for the measures of anti-gay and anti-Semitic hate crimes were they able to find slight significant effects for only one of the methods employed for statistical analysis. The authors concluded that given the failure to find a substantive correlation between economic conditions in both their re-analysis of the Hovland and Sears (1940) data and their contemporary study, the application of the frustration-aggression hypothesis to economic conditions and hate crimes is erroneous. Green, Glaser, and Rich suggest that the frustration-aggression hypothesis is a short lived process that dissolves over time, and cannot be attributed to long-term frustrations resulting in aggressive acts such as hate crimes. Other theories have attempted to explain the occurrence of hate crimes in relation to the risk factors found within the community. One of these theories is the realistic group conflict theory. This theory arose from a series of experiments conducted by Sherif and Sherif in 1953 (cited in Jackson, 1993). The research was conducted on young boys at summer camps. First, Sherif and Sherif created group conflict through competitive activities. They then gave the groups of boys’ communal goals to work on which eventually led to cooperation. Jackson (1993) summarized Sharif’s theory as follows: “intergroup hostility is produced by the existence of conflicting goals (i.e., competition) Community Risk Factors 17 and reduced by the existence of mutually desired superordinate goals attainable only through intergroup competition.” A related theory to the realistic group conflict theory is the group position theory, which emphasizes the distinct hierarchical positions of racial groups in relation to intergroup conflict (Bobo, 1997). Proponents of the group position theory argue that the realistic group conflict theory is too dependent on the material causes of conflict and not the effect of overreaching relative positions between different racial and ethnic groups. Threat between racial/ethnic groups arises out of a weakening of the dominant position of one group over another group which might arise out of changes in economics or other competitions over valued resources. Several theories have further attempted to explain the intergroup hostility that occurs when race/ethnic populations change. The power-threat hypothesis suggests that as the minority population increases, the number of hate crimes against the minority also increases (Tonlay, Beck, & Massey, 1989 cited in Green, Strolovitch, & Wong, 1998). The White majority begins to feel threatened by the influx of minorities and attacks the minorities in response to this threat. An alternate explanation is the power-differential hypothesis, which claims that hate crimes targeting the minority are greatest when their numbers are smaller (Levine & Campbell, 1972, cited in Green, Strolovitch, & Wong, 1998). The White majority is more at liberty to attack minorities whose smaller numbers inhibit their ability to defend themselves. The defended neighborhood hypothesis similarly believes that neighborhoods with a large white majority will have a greater number of hate crimes, particularly when the minority population is increasing (Green, Strolovitch, & Wong, 1998). Suttles (cited in DeSena, 1990) defined the defended Community Risk Factors 18 neighborhood as “the residential group which seals itself off through the efforts of delinquent gangs, by restrictive covenants, by sharp boundaries, or by a forbidding reputation.” Hate crimes are one of the more harmful ways in which a neighborhood (or more accurately, a select few in the neighborhood) attempts to protect itself from the perceived threat of foreign invasion. Corzine, Creech and Corzine (1983) tested the power-threat hypothesis by looking at the incidence of lynchings in the South in comparison to the concentration of African Americans. The power-threat hypothesis predicts that the incidence of African American lynchings will be positively related to higher concentrations of African Americans in the same area. The authors also wished to test a previous finding that data in support of the power-threat hypothesis is only found in the Deep South and not elsewhere. Corzine et al. gathered their data on lynchings occurring between 1889 and 1931 from the National Association for the Advancement of Colored People and compared the incidence of lynchings to the percentage of African American in the corresponding counties during the same time period. The data was tested using both cross-sectional and longitudinal analysis. Both of these analyses found a relationship between lynching incidence and African American concentration in the Deep South, but not in the Upper South, supporting Corzine et al.’s hypotheses. While this study only provides support for a specific type of hate crime and offenders/victims occurring in a very specific time and place, it nonetheless calls attention to the need for further studies examining a broader range of hate crime occurrences and race/ethnic populations. Another study on lynchings in the South at the turn of the century looked at the spatial effects of the lynchings themselves (Tolnay, Deane, & Beck, 1996). The Community Risk Factors 19 researchers tested the contagion model, which predicts an increase in lynchings in areas surrounding a lynching occurrence, and the deterrence model, which predicts that the occurrence of a lynching in one area will deter lynching in the surrounding area. Lynchings between the time periods 1895-1899, 1905-1909, and 1915-1919 were used to “determine the form and assess the magnitude of the spatial dependence of lynchings” (Tolnay, Deane, & Beck, 1996). The researchers controlled for the concentration of African Americans, as well as socioeconomic, cultural and geographic factors, and lynching history. Their analysis found a negative relationship between lynching incidence and the surrounding level of lynchings, supporting the deterrence model. Tolnay et al. concluded that one possible explanation is that Whites believed that neighboring lynchings would sufficiently deter African Americans from what they believed to be offensive behavior requiring lynching. They also supposed that African Americans may have reacted to neighboring lynchings by attempting to avoid negative situations which might lead to lynching. This study demonstrates that other explanations beyond economic and ethnic factors may account for lynching incidence. Green, Strolovitch, and Wong (1998) studied the relationship between race/ethnic populations and hate crime incidence in order to test the various theories on race/ethnic change and intergroup violence presented above. Specifically, the authors asked “to what extent does minority victimization depend on the ways in which the proportions of different racial groups have changed over time?” Green and his colleagues also tested the relationship between economic factors and hate crimes in this study. Their sample consisted of hate crimes reported to the Bias Crime Unit in New York City between 1987 and 1995. The researchers measured racial/ethnic change by comparing the 1980 and Community Risk Factors 2