Pages from J158v05n01-2.pdf (2005) Journal of Forensic Psychology Practice, Vol. 5(1) ___________________________________ AREA REVIEW ______________________________ Assessment of Hate Crime Offenders: The Role of Bias Intent in Examining Violence Risk Edward Dunbar, PhD Jary Quinones, PhD Desiree A. Crevecoeur, PhD ABSTRACT. This study investigated the criminal histories and violence risk of a sample of 204 hate crime offenders. Record review of the offender’s criminal history was rated on the HCR-20 and Cormier-Lang scales. Crime reports were rated for the severity of the bias offense on the Victim Functional Impact scale and the offender’s targeting of outgroup victims–i.e., the bias intent. Findings indicated that 56% of the offenders had prior criminal convictions; HCR-20 ratings were comparable to those found in other offender groups and were correlated with ------------------------------------------------------------------------------------------------------------- Edward Dunbar is affiliated with Pacific Psychological Associates. Jary Quinones is a PhD candidate affiliated with the Los Angeles Police Department. Desiree A. Crevecoeur is a PhD candidate affiliated with the University of California Los Angeles. Address correspondence to: Edward Dunbar, Department of Psychology, University of California Los Angeles, Los Angeles,CA90024USA(E-mail: edunbar@ucla. edu). Journal of Forensic Psychology Practice, Vol. 5(1) 2005 http://www.haworthpress.com/web/JFPP © 2005 by The Haworth Press, Inc. All rights reserved. Digital Object Identifier: 10.1300/J158v05n01_01 1 http://www.haworthpress.com/web/JFPP the severity of the hate crime. The severity of the criminal history as measured on the Cormier-Lang scale, the number of prior arrests, and number of criminal convictions were significantly greater for offenders who targeted racial minority victims. Offenders who belonged to bias oriented groups had more extensive and violent criminal histories and committed more severe hate crimes. Findings are considered in terms of clinical intervention and risk assessment practices with hate crime of- fenders. [Article copies available for a fee from The Haworth Document Delivery Service: 1-800-HAWORTH. E-mail address: Website: © 2005 by The Haworth Press, Inc. All rights reserved.] KEYWORDS. Hate crimes, risk assessment, bias motivation In the U.S., a crime that is motivated in whole or in part by the offend- ers animus towards their victims’ social ingroup may be prosecuted un- der either state or federal statutes (Levin, 1999). Since the passage of the Federal Hate Crime reporting act in 1990, bias motivated crimes have been monitored nationally by the criminal justice system. This has spurred research concerning hate crime victimization (Ehrlich, 1992; Herek, Gillis, Cogan, & Glunt, 1996). However, little attention has been given to the offenders of bias motivated crimes. The assessment of these offenders is particularly important given that the majority of hate crimes are not committed for material gain (e.g., robbery) or as a consequence of reactive revenge-type aggression as is found in domestic or work- place violence (Sullaway, in press). The absence of information for this offender group has led to problems in the identification (Jacobs & Pot- ter, 1998) and prosecution (Dunbar, 1999) of suspects of hate crimes. To date there is no available information concerning the risk for vio- lence posed by hate crime offenders. The popular stereotypes of perpe- trators of hate crimes as skinhead youth or psychologically disturbed members of hate groups has not aided criminal justice and mental health professionals in their rehabilitation efforts with this special class of of- fender. Research is needed that addresses the base rates of criminality and violence risk of hate crime offenders. This would contribute to the practical issues of case management and treatment as well as the gener- ation of theoretical questions concerning risk for commission of hate crimes. Without such information it is very difficult to determine whether hate crime offenders constitute a relatively greater risk for 2 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE http://www.HaworthPress.com recidivistic violence and, concomitantly, are the intervention needs of this offender group. Hate crimes must evidence a clear intent to engage in violence against victims of a specific social group. Accordingly, differences in the offender’s bias intent may be important to the issue of risk assess- ment. This study examined the issue of violence risk with hate crime of- fenders. This was done by investigating the base rates of prior criminal activity and assessing their risk for violent recidivism. This study also considered whether the offenders’ bias intent–as evidenced by the so- cial group of their victim–was related to differences in violence risk and prior criminal activity. VIOLENCE RISK ASSESSMENT WITH HATE CRIME OFFENDERS An important question in the area of hate crime research concerns whether hate crime offenders pose a significant risk for violence. It is not known if offenders of hate crimes frequently perpetrate non-bias motivated crimes and if so, the seriousness of their criminal activity. Additionally determination of the risk for violent recidivism of hate crime offenders needs to be examined vis-à-vis other offender groups (Boer, Wilson, Gauthier, & Hart, 1997; Steadman et al., 2000). Empiri- cally-derived risk assessment incorporates behavioral and development predictors of violence (Menzies, Webster, McMain, Staley, & Scag- lione, 1994). These assessment tools seek to provide a systematic means by which to evaluate violence risk. However, historical, that is static, risk indicators fail to incorporate dynamic–i.e., changeable–elements of risk. Webster, Douglas, Eaves, and Hart (1997) have sought to remedy this problem by incorporating both historical and dynamic elements in the assessment methodology of the HCR-20. Determining risk for vio- lence therefore ought to consider both fixed indicators of risk as well as changeable clinical and contextual characteristics of the offender. A question in the study of violence research that has received little at- tention concerns whether risk estimates are related to the severity of subsequent violent crimes (Clear, 1988). Might the criteria which indi- cate risk for future violence additionally provide information on how vi- olent the hate crime is itself? In other words, do indicators of risk for violence also demonstrate a relationship to the severity of the subse- quent offense, and not simply whether it is more or less likely to occur? Area Review 3 THE QUESTION OF BIAS INTENT AND RISK FOR VIOLENCE By definition hate crimes reflect a discernible intent of the offender to target victims of a specific social outgroup. Intent is an important ele- ment in criminal law (American Law Institute, 1962; Black, 1979). Hate crime offenders intentionally and consciously commit crimes that target members of specific outgroups. Sullaway (in press) has noted that in hate crimes, intent reflects what the offender perceives as the good that will come from their actions; this may include inciting racial unrest in a community, forcing ethnic and sexual minorities out of a neighbor- hood by intimidation (Umemoto & Mikami, 2000), or targeting crime victims–gay men or undocumented persons, for example–who are less likely to seek the assistance of law enforcement personnel. In the case of hate-oriented criminal groups, intent may include efforts to force the out-migration of foreign-born persons from nation states, as well (Watts, 2001). Differences in the bias intent of these offenders may therefore be important in terms of risk assessment. As Sullaway (in press) has noted, “psychology can make a contribution to the develop- ment and refinement of techniques to collect evidence that best captures the various motivations involved in hate crimes.” Dunbar (2003), in a record review of 58 convicted hate crime offenders, found that ratings of a more discernible bias motivation were related to commission of in- strumental violence and an intent to target racial minorities. This find- ing suggests that within-group differences of bias intent may be worth consideration in the assessment and treatment of hate crime offenders. The issue of bias intent has also been considered in terms of societal norms that condone violence against gays, lesbians, and bisexuals. Herek (2000) and Franklin (2000) have suggested that bias against gay men and lesbians constitutes a cultural norm of U.S. society and is not per se deviant from mainstream values. Franklin’s (2000) research with college-age students suggests that harassment and aggression targeting gays and lesbians is quite frequent, and by implication engaged in by otherwise non-criminal individuals. Accordingly, there may be discern- ible offender differences in the risk for violence, if indeed aggression against gays and lesbians is perpetrated by more socially conventional offenders when compared to offenders of religious or ethnic/racial hate crimes. Another potential factor in the assessment of violence risk with this population concerns offenders who associate with other bias-oriented individuals. Members of both hate-oriented groups and members of 4 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE racialized criminal gangs who engage in hate crimes–i.e., the commis- sion of violence that is not better accounted for by inter-gang conflict but rather the targeting of members of ethnic outgroups–may constitute an important sub-group of hate crime offenders. Offenders who demon- strate a discernible bias motivation, particularly when it results in their involvement with other hate-oriented group members, may engage in distinctly different types of bias motivated violence. As suggested by Ezekiel (1995), attention needs to be given to offenders who belong to racialized or hate-oriented groups which espouse a hate-based ideology may pose a particular risk for recurring hate violence. This study examined the following research questions: (1) what are the base rates for criminal activity–i.e., presence of criminal history–for hate crime offenders? (2) what is the risk for violence of these offenders vis-à-vis other criminal samples? (3) does the seriousness of the of- fender’s history for violence vary based upon their bias intent, as deter- mined via the targeted outgroup? and (4) are empirically-derived risk assessment systems such as the HCR-20 correlated with the severity of violence reported in the hate crime? METHOD Participants Review of 814 reported hate crime offenses during the calendar year of 1999 in a large metropolitan area were reviewed in the current study. The cases were collected by an investigative unit of a law enforcement agency in a community located in the Western United States. Of these 814 cases, demographic information on the offender was available for 581 (71%) of the offenses. A subset of these cases (n = 204) resulted in the apprehension and identification of a suspect of the hate offense. The criminal histories of these 204 hate crime offenders were then reviewed and analyzed. Materials Offender Demographic Characteristics. Information from the crime report was coded for the offender’s age, gender, and ethnicity/race. In addition, the identified bias motivation, reported by the responding offi- cer was also coded in our record review. Area Review 5 Offender’s Criminal History. The criminal histories of the identified bias offenders were tabulated through analysis of state and federal data management systems. Information analyzed included the specific charges that led to arrest or conviction for adolescent and adult offenses in the state of California and in the U.S., motor vehicle infractions, open and closed probation activity, bench warrants issued, probationary vio- lations, and sentencing for federal parole. HCR-20 (Webster, Douglas, Eaves, & Hart, 1997). The HCR-20 is a structured clinical guide, comprised of risk factors for violent behavior. The 20 rating items are classified into a 10-item Historical (H), a 5-item Clinical (C), and a 5-item Risk Management (R) scale. These 20 dimen- sions have been found in prior research to effectively estimate risk of fu- ture violence. The H criteria include ratings for prior violent and risk activities, the C criteria define the attitudes and clinical status of the of- fender at time of the evaluation, and the R criteria reflect the presence of destabilizing factors related to violent recidivism. Each item is scored by the examiner on a 3-point scale to denote if the specific variable (e.g., his- tory of previous violence) is clearly present, somewhat present, or absent (scored 2, 1, and 0, respectively). The HCR-20 integrates static historical indicators with dynamic indicators of risk for future violence. The HCR-20 has been used in studies with offender groups with mixed psy- chiatric diagnoses (Douglas, Klassen, Ross, Hart, & Webster, 1998), yielding an H score of 12.45 (SD = 3.6), C score of 5.47 (SD = 2.46), and R score of 6.64 (SD = 2.3). In the Dunbar (2003) study with convicted hate crime offenders, the HCR scale values were 9.35 (SD = 5.72) for the historical criteria, 6.11 (SD = 1.85) for the clinical criteria, and 6.77 (SD = 2.96) for the risk management criteria. Cormier-Lang Crime Index (Quincey, Harris, Rice & Cormier, 1998). This rating scale is based upon an earlier system developed by Akman and Normandeau (1967) to quantify the severity of the offender’s crimi- nal history. Quincey et al. (1998) note that “this system can be used when only official police ‘rap sheet’ information is available” (p. 250). The rat- ing system is organized into Class One offenses, consisting of 16 individ- ual weighted offense criteria (e.g., assault causing great bodily harm = 5), representing sub-types of violent crimes and Class Two offenses consists of 21 weighted items for non-violent crimes (e.g., theft under $500 = 1; extortion = 5). A total score for all offenses can also be computed. With a sample of 58 convicted hate crime offenders (Dunbar, 2003), the mean for the total Cormier-Lang scales was 15.40 (SD = 18.92). Victim Functional Impact. An estimate of the crime’s impact upon daily life functioning for the victim was determined via a criterion rat- 6 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE ing method. The professional staff (n = 22) of the Victim Witness Assis- tance Program of the Los Angeles County District Attorney’s Office was surveyed to provide a severity of impact rating for the 18 acts of ag- gression noted in the crime reports. These individuals provide direct hu- man service consultation and medical referral to the Victims of Violent Crimes for Los Angeles County. Typically, they interact with clients dur- ing the acute post-event phase of crime victimization. Each act of aggres- sion was rated on a Likert scale, as derived from the Axis 4 rating system for severity of psychosocial stress, as first proposed in the Diagnostic and Statistical Manual-Third Edition, Revised (DSM-III-R) classification sys- tem (APA, 1988). The rating range for severity of functional impact for each individual act of aggression ranged from 60 (catastrophic) to 10 (not at all). This rating methodology provided an estimate of the victim’s post-event impairment, based upon the type of aggression reported in the crime report. For example, assault with a deadly weapon had a rating value of 52.23 (falling in the range of catastrophic-to-extreme), physi- cal assault (extreme-to-severe impact) had a mean of 45.47, whereas be- ing the target of printed humor or hate speech had an impact value of 27.28 (moderate-to-mild impact). Using these rating values, an estimate of post-event victim impact was derived. This was determined by ag- gregating the individual acts of aggression identified for the offense, as reported in the crime report. Ratings on the Functional Victim Impact scale produced a mean of 69 (SD = 21) for 1,538 hate crimes as deter- mined via record review with a mean inter-class correlation of .76 for the rating items (Dunbar, 1997). PROCEDURE The methodology incorporated record review and content analysis of the criminal history and the bias (index) crime. Classification of the in- dex offense as being bias motivated employed a multi-step process. First, the index offense had to be identified as bias motivated by the re- sponding officer. This crime report was then reviewed by a precinct hate crime coordinator, who affirmed that the offense met the legal stan- dards to be classified as bias motivated. The crime report was then for- warded to a special investigation unit, where the report was again reviewed to confirm or deny that the index offense met the standards to be reported as a hate crime. This process allowed for the removal of of- fenses that did not meet the legal standard of being both a crime and as being bias-motivated. Information on the index crime was recorded by Area Review 7 the responding officer for bias crime motive (e.g., race/ethnicity, reli- gion), number of offenders involved in the crime, and demographic characteristics (gender, race, and age) of the perpetrator(s). Analysis of the crime reports for the bias offense was conducted by a team of university research assistants under the supervision of the first author. The reports were examined to determine the specific bias intent (religion, sexual orientation, for example) of the crime. Offender char- acteristics of age, race/ethnicity, and gender were also coded. Offenders identified on the crime reports as having been members of hate oriented gangs and groups were also noted. The characteristics of the index crime were rated in terms of a variety of behavioral characteristics, which were used to compute the estimated crime severity on the Victim Functional Impact scale. The first author provided training and consul- tation in the coding of the Victim Functional Impact scale. The criminal histories of the identified hate crime offenders were rated via record review for all crime activity prior to commission of the hate crime. Information included the identified offender’s “rap sheet” of prior state and federal arrests and convictions. The criminal history was searched based on the child/adolescent and adult history of arrests and convictions, the motor vehicle record, and federal crime information. The initial criminal record search was conducted by the second author. The criminal history was then reviewed to establish severity ratings on the Cormier-Lang scale. Content analysis of the criminal records was used for scoring the HCR-20. This was coded in accordance with the rating criteria employed in the respective scoring manuals. These were coded by the first and third authors. The evaluation of the criminal his- tories and assignment of levels of violence risk was done without re- view or knowledge of the index crime, other than that the crime had been classified as being a hate crime. RESULTS The demographic characteristics of the 581 hate crime offenders in- cluded 497 (85.5%) men, 82 (14.1%) women, and 2 offenses were com- mitted by men and women together (.4%). The race/ethnicity of the 581 bias crime offenders included 17.8% African American, 2.3% Asian- Pacific, 44.8% Euro-White, 30.4% Latino, and 4.7% of diverse ethnic and/or multi-racial backgrounds. The distribution of the offender’s race and ethnicity is approximate to that of the demographic representation of the metropolitan area in which the crimes occurred. 8 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE For the 204 identified hate crime offenders 87% were men and 13% were women. The mean age of the identified bias offenders was 32.69 (SD = 14.04, range = 12 to 81); 84% of the offender sample were adults at the time the crime was committed. The race/ethnicity of this group in- cluded 15% African American, 2.6% Asian-Pacific, 48% Euro-White, 26.2% Latino and 7.2% of diverse or multiracial backgrounds. The of- fenses committed by these identified offenders revealed that 57.6% of the offenses were crimes again the person, 17.9% were property crimes, and 24.5% were verbal threats of harm to the person. There were no signifi- cant differences in the demographic representation of the sample of 581 hate crime offenders and the group of 204 identified (i.e., apprehended) offenders from which they were drawn. Ratings from the crime reports of the hate (index) offense indicated that 16% (n = 31) of the identified of- fenders were members of a hate-oriented criminal gang or group. Base rates were computed for history of arrests and convictions (re- search question one). At the time of the commission of the bias crime, 59% of the 204 identified offenders had a prior history for arrest (M = 4.04, SD = 7.58) with 58 % having had a prior criminal conviction (M = 3.22, SD = 5.98); 33% had more than one prior conviction. For the identi- fied offenders the Cormier-Lang estimates for severity of history for vio- lent offenses was 8.28 (SD = 15.97; range of 0 to 101) and 6.71 (SD = 10.49; range = 0 to 50) for non-violent offenses. Estimation of risk for violence (research question two) was assessed on the HCR-20. For the identified bias offenders, the HCR-20 Histori- cal indicator (H) scale mean was 8.21 (SD = 4.74; alpha reliability coef- ficient = .94), the Clinical (C) risk mean score was 5.39 (SD = 2.46; alpha reliability coefficient = .90) and the Risk Management (R) mean was 5.53 (SD = 2.97; alpha reliability coefficient = .95). The computed mean Co- hen kappa for inter-rater agreement on the HCR-20 was .58 (range of .90 to .33). Cichetti and Sparrow (1981) note that kappa values of .40 to .59 are fair and values from .60 to .74 are considered good. The HCR-20 scores are similar to a sample of 175 offenders with mixed psychiatric di- agnoses (Douglas et al., 1998). The total number of prior arrests and con- victions were significantly correlated with the HCR-20 Historical (r = .60, p < .001, r = .67, p < .001), Clinical (r = .53, p < .001, r = .48, p < .001), Risk Management (r = .55, p < .001, r = .62, p < .001) scale scores. The HCR-20 risk indicators were also correlated with the Cormier-Lang scales. The HCR-20 Historical (i.e., static) indicators were correlated to severity of prior violent crimes (r = .64, p < .001) and non-violent crimes (r = .66, p < .001). The dynamic (i.e., changeable) Clinical (r = .49, p < .001; r = .57, p < .001) and Risk Management (r = .56, p < .001; r = .63, Area Review 9 p < .001) indicators were also correlated with the estimates on the Cormier-Lang for severity of prior violent criminality and non-violent criminality, respectively. These findings are presented in Table 1. The criminal records revealed a variety of historical indicators re- lated to heightened risk for violence. Forty-eight percent of the offend- ers had a prior arrest or conviction for a violent crime; 36% had an offense for violence prior to the age of 20. Twenty-three percent of the sample had a criminal record for substance abuse. A quarter (26%) of the identified bias offenders had family histories marked by parental separation and/or domestic violence; these were typically noted in the criminal record as having had “unfit” childhoods. That is, under the criminal code, the offender required institutional intervention as a mi- nor child. Thirty-eight percent of the offenders had histories which im- plicated significant occupational problems; 45% had had prior super- vision failures (e.g., probation and parole violations) with the criminal justice system. 10 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE TABLE 1. Criminal History, HCR-20, and Cormier-Lang Risk Indicators of 204 Identified Bias Offenders Percent with one or more offense M SD Range Criminal History Misdemeanor arrests Misdemeanor convictions 42% 45% 1.37 1.29 2.57 2.53 0-18 0-18 Felony arrests Felony convictions 44% 36% 2.04 1.32 3.99 2.57 0-28 0-13 Juvenile convictions 22% .78 2.15 0-15 Prior probation sentencing Probation at time of bias crime 28% 31% .64 .65 0-7 0-1 Motor vehicle infractions Bench warrants issued 32% 14% 1.02 .30 2.43 .98 0-18 0-8 HCR-20 Historical Indicators 8.21 4.74 0-19 Clinical Indicators 5.39 2.46 0-10 Rick Management Indicators 5.53 2.97 0-10 Cormier-Lang: Severity of Criminal History Violent Offenses 8.28 15.97 0-101 Non-Violent Offenses 6.71 10.49 0-50 Research question three examined whether the offender’s bias intent was related to differences in the risk for violence and extensiveness of the criminal history. The bias intent classified in the crime reports in- cluded 116 (54.4%) due to racial or ethnic bias, 25 (11.3%) due to reli- gious bias, 56 (26.5%) of the hate crimes were motivated by the victim’s sexual orientation and 6 (2.9%) were gender bias crimes; one case could not be classified for any of the identified categories. A series of One- Way Analysis of Variance tests were computed for the three primary categories for bias intent (race/ethnicity, religion, and sexual orienta- tion) by the HCR-20 scores, aggregate history of arrests and convic- tions, and the Cormier-Lang estimates of the severity of the criminal history for violent and non-violent offenses. Results revealed that dif- ferences in the offenders’ targeted outgroup were significantly related to the number of prior arrests (F = 3.62, p < .05) and number of convic- tions prior to commission of the hate crime (F = 4.38, p < .01). In both in- stances, Scheffe contrasts revealed that race/ethnic hate crimes were related to more extensive criminal histories than offenders who commit- ted hate crimes due to religious bias. ANOVA results for the Cormier- Lang estimates of the severity of the history of violent offenses by bias in- tent was also significant (F = 4.39, p < .01), with offenders who targeted racial and ethnic minorities having significantly more severe histories of violence than offenders who committed hate crimes due to religious or sexual orientation bias. Table 2 summarizes these findings. A separate series of analyses were computed with bias offenders who were classified on the crime reports as having been a member of a hate-oriented group or racialized criminal gang. When compared to other offenders in the sample, hate gang members had higher scores for total arrests (t = 3.03, p < .003) and total convictions (t = 2.62, p < .008), as well as higher HCR-20 Historical (t = 3.41, p < .001), Clinical (t = 2.01, p < .01), and Risk Management (t = 4.91, p < .001) scores. The hate gang members also had higher Cormier-Lang scores for the severity of violent criminal history (t = 2.12, p < .04) and non-violent criminal history (t = 2.13, p < .03); they also perpetrated significantly more severe hate crimes as measured on the Victim Functional Impact scale (t = 5.71, p < .001). Research question four examined whether empirically-derived risk ratings demonstrated a relationship to the severity of the hate crime of- fense. For the 204 identified bias offenders, the mean Functional Victim Impact score was 93.67 (SD = 28.39; range from 27 to 204) in the com- mission of the hate crime. The HCR-20 Historical (r = . 27, p < .001) Clinical ( r = . 33, p < .001) and Risk Management (r = . 31, p < .001) scale scores were all correlated to the Victim Functional Impact scale. Area Review 11 The Victim Functional Impact score was not significantly correlated to the Cormier-Lang scores for severity of the criminal history or the total number of prior convictions or arrests. DISCUSSION This study examined the issue of risk assessment with hate crime of- fenders from three different perspectives. This included determination of the offender’s criminal history, estimation of the severity of prior of- fenses, and assessment of the relative risk for future violence. Collec- tively findings indicated that the majority of identified hate crime offenders had a prior criminal history and that differences in this of- fender group were related to the targeted outgroup of their victims. Hate crime offenders who committed offense based upon racial ani- mus had more extensive and violent criminal histories. By contrast, of- fenders who evidenced intent to target religious outgroups had fewer prior offenses and less severe criminal histories. Offenders who tar- geted gay and lesbian victims fell in a middle range between these two other offender groups. Record review indicated that many offenders 12 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE TABLE 2. Identified Bias Offenders Risk Indicators by Bias Intent Racial/Ethnic (n = 116) Sexual Orientation (n = 56) Religion (n = 25) F Criminal History Total Arrests 5.51 2.59 1.17 3.62* Criminal History Total Convictions 4.23 1.95 1.77 4.38** HCR-20 Historical Indicators 8.80 7.74 6.64 1.23 HCR-20 Clinical Indicators 5.48 5.23 5.41 .09 HCR-20 Risk Mgmt. Indicators 5.88 4.81 4.33 1.02 Cormier-Lang Aggregate History Violent Offenses 11.21 4.03 2.81 4.39** Cormier-Lang Aggregate History Non-Violent Offenses 6.90 4.02 3.75 1.11 *p < .05, **p < .01 who targeted gay and lesbian victims had a prior history of violence which was not bias motivated. Thus, while offenders of hate crimes tar- geting sexual minorities had less extensive criminal histories than many other (i.e., racially motivated) hate crime offenders, they were likely to have evidenced prior anti-social behavior. These findings do not, how- ever, conclusively address the proposition of Franklin (2000) that these offenses are committed by individuals with less extensive criminal his- tories when compared to a general criminal population. Our findings also indicated that the members of hate-oriented gangs evidenced significantly greater risk for violence in general and indeed had more extensive criminal histories than other hate crime offenders. They were more violent by their criminal histories, they evidenced more historical and dynamic risk for on-going violent behavior, and they en- gaged in significantly more violent forms of aggression in the commis- sion of the hate crime. This small group of individuals indeed reflects the popular image of the violent hate crime offender as a highly disso- cial and aggressive individual. However, our findings underscore that these offenders constitute a very small percentage of individuals who commit hate crimes. Implications for Clinical Assessment The findings of the risk assessment criteria suggest that hate crime offenders pose treatment and probationary challenges comparable to that found for mentally disordered offenders. The implication of our findings is that hate crime offenders frequently pose a risk for violence based upon both static and dynamic factors. This and the prior study of Dunbar (2003) both found that the static historical risk ratings on the HCR-20 were modestly lower when compared to published data with the measure for other offender groups. By comparison both of the stud- ies using the HCR-20 with hate crime offenders yielded moderately higher estimates for dynamic clinical and risk management factors. That is, prior life circumstance related to violence risk were modestly lower when compared to other samples using the HCR-20 criteria, but revealed somewhat greater risk for violence that was contemporaneous to the commission of the hate crime. The information from our sample indicated many–though by no means all–of these offenders evidenced problems frequently found in patients diagnosed with a personality disorder–i.e., occupational prob- lems, relationship failure, code-breaking behaviors, and impulsivity. The relationship between personality disorders–specifically narcis- Area Review 13 sism–and pathological forms of bias has been discussed by Bell (1979;1978). Pathological forms of bias have been identified as an area warranting further investigation by the DSM-V workgroup (Kupfer, First, & Regier, 2002). At present, however, a program of research has not been established to examine the relationship, if any, between per- sonality disorders and intergroup aggression, yet alone commission of hate crimes. Independent of the serious problems of daily living found with these offenders is their manifestation of aggressive intergroup behaviors. As Dunbar (2003) has suggested, recidivism risk for bias offenders may be further complicated when a discernible hate ideology is also present. These two components–a heightened risk for violence and a discernible hate ideology–may incrementally increase the probability for recidi- vistic bias criminality. Further study of the role of bias motivation as an independent risk factor for this group of offenders is an area in need of examination. The assessment of hate crime offenders needs to be considered in terms of the debate over the relative merits of clinical and psychometric methods versus empirically-derived or “file drawer” approaches to vio- lence risk (Rogers, 2000; Hilton & Simons, 2001). We would hope that both informed clinical opinion and the use of empirically-derived infer- ences will be used in the evaluation of this special class of offender. Re- search-driven methods optimally utilize clinical knowledge in risk determination, even when psychometric measures and clinical inter- views are not part of the assessment methodology. With unique sam- ples, such as offenders of hate crimes, a multi-faceted approach to assessment can provide greater specificity in risk determination. In ad- dition, our findings suggested the importance of within-group factors of bias intent–as evidenced by the offender’s targeting of specific out- group victims–in risk assessment. The relationship between the HCR-20 ratings and the severity of the subsequent hate crime (as determined by a separate rater group) points to other useful applications when historical, clinical, and contextual in- dicators of violence risk are integrated. We found that the aggregated crime histories and the estimated valuations of the severity of the prior criminal offenses were unrelated to the severity of the hate crime itself. This suggests that risk assessment of violence based solely upon the criminal history, even for offenders with extensive histories of violent crimes, may be less clearly related to the severity of future violence than what can be deduced through the use of an empirically-derived rating system. Our findings illustrate how these methods can be employed to 14 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE identify specific treatment goals (e.g., substance abuse, impulsivity) in the rehabilitation of bias motivated offenders. This is particularly im- portant given that many stakeholder groups involved in responding to bias crimes have limited knowledge of what factors heighten risk for vi- olent recidivism. The utilization of a risk assessment strategy that em- ploys empirically-derived indicators along with crime investigation information can inform both law enforcement professionals and policy makers alike. The violence risk posed by hate crime offenders extends to commu- nities as well as individual victims. Given that hate crime offenders en- gage in violence against social groups, the magnitude of risk is more complex. Bias motivated violence as Bell (2003) has observed, consti- tutes a public health risk. When considered in this light, offenders of hate crimes, like serial rapists, constitute a risk to the community which is dif- ferent by kind when compared to many violent offenders. As such, hate crime offenders constitute an additional societal risk, one that may inflict a substantial social psychological injury to both vulnerable individuals and groups. At present, there is no agreed-upon standard or criteria by which to assign risk to the community, in terms of individual offenders of hate crimes. This is an area in need of future investigation. Implications for Intervention These findings point to the need to reduce identified risk factors of vi- olence for hate crime offenders such as poverty, occupational failure, and problematic familial relationships, as much as to address their cul- tural misperceptions. Comprehensive wrap-around interventions are needed to reduce the recidivistic behaviors of bias motivated offenders. Interventions such as Multi-Systemic Therapy (Henggeller, Melton, Brondino, Scherer, & Hanley, 1997) which have proven effective with young at-risk offenders provide one such example of a viable treatment model. Likewise, innovative criminal justice practices such as drug court programs and specialized probationary monitoring need to be im- ported to the area of hate crime offender management. To date, how- ever, interventions to ameliorate an offender’s intergroup bias have not been shown to contribute to the reduction in re-offense rates or desistance of at-risk behavior for recidivism. We feel that comprehen- sive risk assessment is needed in order to determine the treatment needs of bias motivated offenders. The current findings underscore the serious risk posed by hate crime offenders, one which may be obfuscated by the more controversial issue Area Review 15 of how “criminal” hate crimes really are. This controversy, and on occa- sion dismissal, of bias crimes as not being “real crimes” as Boyd, Berk, and Hammer (1996) note, may lead some law enforcement personnel to not identify the hate element in crimes they report. This may result in the minimization of the seriousness of both the offense and the larger threat posed by the offender to their community. Nolan and Akiyama (1999) note that the effectiveness of law enforcement agencies to en- force hate crime laws is influenced by organizational attitudes and com- mitment, community relations, economic resources, and the individual beliefs of law enforcement personnel. The current findings need to be considered in terms of both the com- munity and institutional forces which shape how hates crimes are iden- tified. In the current study, the large metropolitan agency from which the sample was drawn had several structural advantages in the detection of hate crimes which may be absent in other law enforcement organiza- tions. This included a centralized investigative unit responsible for monitoring hate crimes and the coordination at the precinct level with officers who served as point personnel in crime investigation and inter- diction. The metropolitan area additionally was supported by several victim advocacy and human relations groups who monitored law en- forcement responsiveness to hate crimes. In jurisdictions absent these institutional forces, the identification of hate crimes may well be limited to only the most egregious and sensational events, while failing to iden- tify the numerous offenses of verbal threat and property damage which meet the legal standards as being hate crimes. Limitations in the Current Study This study linked record review of criminal activity with information about hate crime perpetration, as found in the crime report. The use of crime scene data to make inferences about the offender is in many cases highly impressionistic. As Alizun, Bennell, Mukros, and Ormoral (2002) note, the derivation of personality characteristics and other intrapsychic states from review of crime scene evidence is highly subjective and absent any discernible methodology. We have tried to avoid these pitfalls by instead examining objective behavioral elements of the crime. All the same, while a great deal can be determined from the criminal record, other sources of information–which were not collected by law enforce- ment personnel–would have provided additional evidence of prior vio- lence and exposure to destabilizers. These dynamic indicators would have further informed our risk determination in terms of both heighten- 16 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE ing risk estimates on the one hand (e.g., living with members of a hate group upon institutional release) and lowering risk concerns on the other (e.g., involvement in 12-step sobriety groups). Research on violence risk assessment has frequently been accom- plished via record review (Harpur, Hare, & Hakstian, 1989). There are numerous studies that have employed ratings based upon record review methods for the classification of violent offenders (Wong, 1988; Harris, Rice, & Cormier, 1991). Our use of the HCR-20 reflects the same limi- tations as found in any record review process. The estimates for vio- lence risk arrived at with the HCR-20 in this study are, at this point, exploratory and need to be validated by future research. There is a need to re-examine our assumptions as to what constitutes risk for bias motivated violence and how to mediate it. Reconsidering the assessment and treatment challenges posed with this offender group would help improve the response of criminal justice and mental health providers to this special class of offender. REFERENCES Akman, D. D., & Normandeau, A. (1967). The measurement of crime and delinquency in Canada: A replication study. British Journal of Criminology, 7, 129-149. American Law Institute (1962). Model Penal Code. Proposed official draft. American Psychiatric Association (1988). Diagnostic and Statistical Manual of Mental Disorders Third Edition–Revised. American Psychiatric Association Press, Wash- ington DC. Bell, C. C. (1978). Racism, narcissism, and integrity. Journal of the National Medical Association, 70 (2), 89-92. Bell, C. C. (1980). Racism: A symptom of the narcissistic personality disorder. Journal of the National Medical Association, 72 (7), 661-665. Bell, C. C. (2003). Racism: Diagnostic and Treatment Considerations, paper. One hun- dred-fifty-sixth Annual Convention of the American Psychiatric Association, San Francisco, CA. Black, H. C. (1979). Black’s law dictionary, fifth edition. St. Paul, MN: West. Boer, D. P., Wilson, R. J., Gauthier, C. M., & Hart, S. D. (1997). Assessing risk for sex- ual violence: Guidelines for clinical practice. In C. D. Webster & M. A. Jackson (Eds.), Impulsivity: Theory, Assessment, and Treatment. New York: Guilford. Boyd, E. A., Berk, R. A., & Hammer, K. M. (1996). Motivated by hatred or prejudice: Categorization of hate-motivated crimes in two police divisions. Law & Society Re- view, 30 (4), 819-850. California State Department of Justice. (1998). Hate Crime in California. Division of Criminal Justice Information Systems. Bureau of Criminal Information Analysis. Area Review 17 Cicchetti D. V., & Sparrow, S. S. (1981). Developing criteria for establishing the interrater reliability of specific items in a given inventory. American Journal of Mental Deficiency, 86, 127-137. Clear, T. (1988), Statistical Prediction in Corrections, Forum: Research in Corrections, 1, 1-40. Douglas, K. S., Klassen, C., Ross, D., Hart, S. D., & Webster, C. D. (1998). Psychometric Properties of HCR-20 Violence Assessment Scheme in Insanity Acquittees. Paper pre- sented at the 103rd annual American Psychological Association, San Francisco, CA. Dunbar, E. (1997). Hate Crimes in Los Angeles County: Perpetration and Victimology Trends. Paper presented at the American Psychological Association Congressional Briefing: “Hate Crimes: Causes, Consequences, and Current Policy. What Does So- cial Science Tell Us?” Washington, DC. Dunbar, E. (1999). Defending the indefensible: A critique and analysis of psycholegal defense arguments of hate crime perpetrators. Journal of Contemporary Criminal Justice, 15 (1), 64-78. Dunbar, E. (2003). Symbolic, relational, and ideological signifiers of bias motivated offenders: Toward a strategy of assessment. American Journal of Orthopsychiatry, 73 (2), 2003. Ehrlich, H. J. (1992). The ecology of anti-gay violence. In Herek G., & Berrill (1992). Hate Crimes: Confronting Violence Against Lesbians and Gay Men. Newbury Park: Sage Publications. Ezekiel, R. S. (1995). The Racist Mind: Portraits of American Neo-Nazis and Klansmen. New York: Viking. Franklin, K. (2000). Antigay behaviors by young adults: Prevalence, patterns, and motivators in a noncriminal population. Journal of Interpersonal Violence, 15, 4. Hare, R. D. (1991). Manual for the Psychopathy Checklist–Revised. Toronto: Multihealth Systems. Harpur, T. J., Hare, R. D., & Hakstian, A. R. (1989). Two-factor conceptualization of psychopathy: Construct validity and assessment implications. Psychological As- sessment, 1 (1), 6-17. Harris, G. T., Rice, M. E., & Cormier, C. A. (1991). Psychopathy and violent recidi- vism. Law and Human Behavior, 15, 625-637. Hart, S. D., Cox, D. N., & Hare, R. D. (1995). Manual for the Psychopathy Checklist: Screening Version. Toronto: Multihealth Systems. Henggeller, S. W., Melton, G. B., Brondino, M. J., Scherer, D. G., & Hanley, J. H. (1997). Multisystemic therapy with violent and chronic juvenile offenders and their families: The role of treatment fidelity in successful dissemination. Journal of Con- sulting & Clinical Psychology, 65, 5, 821-833. Herek, G. M. (2000). The psychology of sexual prejudice. Current Directions in Psy- chological Science, 9 (1), 19-22. Herek, G., Gillis, R., Cogan, J., & Glunt, E. (1997). Hate crime victimization among lesbian, gay, and bisexual adults. Journal of Interpersonal Violence, 12 (2), 195-215. Hilton, N. Z., & Simons, J. L. (2001). The influence of actuarial risk assessment in clin- ical judgments and tribunal decisions about mentally disordered offenders in maxi- mum security. Law and Human Behavior, 25 (4), 393-408. 18 JOURNAL OF FORENSIC PSYCHOLOGY PRACTICE Jacobs, J. B., & Potter, K. (1998). Hate Crimes: Criminal Law and Identity Politics. New York: Oxford University Press. Kupfer, D. J., First, M. B., & Regier, D. A. (2002). A Research Agenda for DSM-V. American Psychiatric Association Press, Washington DC. Levin, B. (1999). Hate crimes: Worse by definition. Journal of Contemporary Criminal Justice, 15 (1), 6-21. Levin, J., & McDevitt, J. (1993). Hate Crimes. New York: Plenum. Menzies, R. J., Webster, C. D., McMain, S., Staley, S., & Scaglione, R. (1994). The di- mensions of dangerousness revisited: Assessing forensic predictions about vio- lence. Law and Human Behavior, 18, 1-28. Nolan, J. J., & Akiyama, Y. (1999). An analysis of factors that affect law enforcement participation in hate crime reporting. Journal of Contemporary Criminal Justice, 15, 111-127. Quincey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. A. (1998). Violent Offenders. Washington, DC: American Psychological Association. Rice, M. (1997). Violent offender research and implications for the criminal justice system. American Psychologist, 52 (4), 414-423. Rogers, R. (2000). The uncritical acceptance of risk assessment in forensic practice. Law and Human Behavior, 24 (5), 595-605. Salekin, R. T., Roger, R., & Sewell, K. W. (1996). A review and meta-analysis of the psychopathy checklist and psychopathy checklist-revised: Predictive validity of dangerousness. Clinical Psychology: Science and Practice, 3, 203-215. Steadman, H. J., Silver, E. Monahan, J., Appelbaum, P., Clark Robbins, P., Mulvey, E. P., Grisso, T., Roth, L. H., & Banks, S. (2000). A classification tree approach to the development of actuarial violence risk assessment tools. Law and Human Be- havior, 24, 83-100. Sullaway, M. E. (in press). Psychological Perspectives on Hate Crime Laws. Psychol- ogy, Public Policy, and Law. Umemoto, K., & Mikami, C. K. (2000). A profile of race-bias hate crimes in Los An- geles County. Western Criminology Review, 2 (Online). Available: http://wcr. sonoma.edu/v2n2/umemoto.html. Watts, M. (2001). Aggressive youth cultures and hate crime: Skinheads and xenopho- bic youth in Germany. American Behavioral Scientist, 45, 600-615. Webster, C. D., Douglas, K. S., Eaves, S., & Hart, S. D. (1997). Manual for HCR-20: Assessing Risk for Violence (version 2). Odessa, FL: Psychological Assessment Re- sources. Wong, S. (1988). Is Hare’s Psychopathy Checklist reliable without the interview? Psy- chological Reports, 62 (3), 931-934. RECEIVED: 07/12/03 REVISED: 12/13/03 ACCEPTED: 12/13/03 Area Review 19 http://wcr