The Enduring Relevance of Race and Ethnicity to Homicide among State Delinquents
- 1. Criminal Justice Studies, Iowa State University, USA
- 2. Department of Psychology, Kent State University, USA
Abstract
In the United States, there are sharp racial and ethnic differences in homicide perpetration, but scholarly disagreement about the sources of those differences. Drawing on a sample of 2,520 male state delinquents selected from a large southern state, the current study examined the role of African American and Hispanic status on homicide offending net the effects of social risk factors and delinquent career. Hierarchical logistic regression models indicated that African Americans and Hispanics had significantly elevated risks for homicide offending. Moreover, the addition of social risk and delinquent career factors reduced the effect for African Americans on homicide from 157% to 123% increased likelihood and reduced the effect for Hispanics from 158% to 92% increased likelihood. Among antisocial samples, African American and Hispanic are at increased risk for homicide offending and these effects withstand the confounding effects of social deprivation and prior antisocial conduct.
Keywords
Homicide; Race; Ethnicity; Delinquency; Criminal careers.
CITATION
DeLisi M, Glotfelty W, Mosley H (2016) The Enduring Relevance of Race and Ethnicity to Homicide among State Delinquents. Ann Forensic Res Anal 3(1): 1026.
INTRODUCTION
Homicide is the most serious criminal offense and it imposes extraordinary costs relating to mortality and human suffering, collateral social burdens, and legal and criminal justice expenditures [1,2]. In the United States of America, homicide offending and homicide victimization are unevenly distributed and disproportionately affect males, youth, antisocial populations, and racial and ethnic minorities. Indeed, diverse data sources have indicated that homicide offending and victimization are dramatically higher among African Americans than whites and moderately higher among Hispanics than non-Hispanic whites [3-7]. These differences are often explained by socioeconomic differences by race and ethnicity and other background risk factors.
The current aim is to examine race and ethnic differences in homicide and to empirically assess whether social risk factors and delinquent career measures ameliorate the increased risk of homicide offending among African Americans and Hispanics.
MATERIALS AND METHODS
Data are derived from a sample (n = 2,520) of adjudicated male delinquents committed to confinement facilities in a large southern state. Information on each juvenile offender was compiled by the juvenile correctional system at a statewide intake unit upon the youth’s commitment and during his institutionalization. All state-committed youth were housed at the intake facility for approximately two months and then were transferred to specific facilities around the state to complete their commitment period. Additional offender data were collected during the youth’s confinement from numerous sources including state- and county-level official records, on-site diagnostic procedures at intake, observations from professional and correctional staff, and self-report information from youth. Data on the official institutional misconduct of state-committed delinquents were collected with standardized instruments at each juvenile facility in the state and maintained at a centralized location.
The sample was comprised mostly by Hispanics (38.3%, n=966), African Americans (35.3%, n=889), and non-Hispanic whites (24.8%, n=625). In the analyses, non-Hispanic whites are the omitted reference group. Two categories of controls were utilized: social risk factors and delinquent career. Social risk factors included seven dichotomous (0=no, 1=yes) family background measures. These were having family members that were in gangs (M= 0.09, SD= 0.29), whether the youth was violent toward his family (M= 0.25, SD= 0.43), family poverty (M= 0.57, SD= 0.50), living in a chaotic home environment (M= 0.69, SD = 0.46), evidence that the youth was physically abused (M= 0.13, SD= 0.34), evidence that the youth was sexually abused (M= 0.11, SD= 0.32), and evidence that the youth was emotionally abused (M= 0.23, SD= 0.42). Basically, these measures were meant to capture the social disadvantage that African American and Hispanic youth disproportionately experience relative to non Hispanic white youth.
Delinquent career included six continuously measured items: age at first state commitment (M=15.3, SD= 1.14, Range= 10-18 years), number of previous out-of-home placements (M= 0.45, SD= 1.12, Range= 0-10), felony adjudications prior to state commitment (M= 1.60, SD= .95, Range= 0-10), and total prior adjudications (M = 1.60, SD = 0.95, Range = 0-10). Gang affiliation is a dichotomous measure (no = 0, yes = 1) of whether the youth is a known gang member. Nearly 36 percent of the sample was gang affiliated at intake. Instant offense gang-related (M = 0.16, SD = 0.37) is a dichotomous measure of whether the youth’s commitment offense was gang-related. These measures were intended to capture core, individual-level effects of criminal propensity or criminality that drive offending careers [8].
Two analytical approaches were used. First, analysis of variance (ANOVA) models were conducted to examine the social risk and delinquent career background factors across the three racial and ethnic groups. Second, hierarchical logistic regression was employed where homicide as commitment offense was regressed on African American and Hispanic status in model 1, social risk factors were added in model 2, and delinquent career measures were added in model 3. The hierarchical approach allows us to examine the effect of African American and Hispanic status on homicide offending while stepping in important covariates. For ease of interpretation, odds ratios and z-scores are presented. Odds ratios can be interpreted as percentage change with the formula ([Odds Ratio-1]*100).
RESULTS AND DISCUSSION
The results in (Table 1) show that all three racial and ethnic groups of serious delinquents had extensive and severe social backgrounds and commensurately severe delinquent careers. Several significant effects were found.
Table 1: ANOVA of Risk Variables across Three Racial/Ethnic Groups. |
|||||
Variable |
African American |
Hispanic |
White |
Omnibus F |
p |
Physical Abuse |
.10 |
.09 |
.23 |
24.76 |
<.0001 |
Sexual Abuse |
.06 |
.08 |
.25a |
56.04 |
<.0001 |
Emotional Abuse |
.20 |
.18 |
.37a |
30.05 |
<.0001 |
Poverty |
.60b |
.63b |
.44 |
22.36 |
<.0001 |
Chaotic Home |
.70b |
.68b |
.73a |
5.86 |
<.001 |
Gang Family |
.08b |
.14b |
.02 |
23.33 |
<.0001 |
Violent Toward Family |
.19 |
.22 |
.39a |
29.40 |
<.0001 |
Age First Commitment |
15.24 |
15.40 |
15.18 |
5.59 |
<.001 |
Prior Felony Adjudications |
1.34b |
1.24 |
1.23 |
7.45 |
<.001 |
Prior Out of Home Placements |
.35 |
.35 |
.79a |
26.80 |
<.0001 |
Gang Affiliation |
.32b |
.52 |
.16 |
83.77 |
<.0001 |
Instant Offense Gang-Related |
.11b |
.28 |
.04a |
69.76 |
<.0001 |
Note: a=white mean is significantly different than African American and Hispanic mean b=African American/Hispanic mean is significantly different than white mean |
White delinquents were significantly more likely than African American and Hispanic delinquents to be emotionally abused, to be sexually abused, and to be violent toward a member of their family. African Americans and Hispanics were more likely to be reared in poverty, to have a chaotic home, and to have gang members in their family. In terms of delinquent career, youths were comparable on onset, African Americans had more felony adjudications, whites had more out of home placements, and Hispanics had greater gang affiliation and greater likelihood that their instant offense was gang-related.
The results in (Table 2) indicate that African American and Hispanic were at robustly elevated risk for homicide offending and while the addition of social risk factors and delinquent career attenuated these relationships, they remained strongly significant.
Table 2: Hierarchical Logistic Regression Model for Homicide as Commitment Offense. |
||||||||
Variable |
Model 1 |
Model 2 |
Model 3 |
|||||
|
OR |
Z |
OR |
Z |
OR |
Z |
||
Race/Ethnicity |
||||||||
African American |
2.57*** |
7.24 |
2.14*** |
5.60 |
2.23*** |
5.69 |
||
Hispanic |
2.58*** |
7.37 |
2.19*** |
5.75 |
1.92*** |
4.47 |
||
Social Risk |
||||||||
Physical Abuse |
- |
- |
1.01 |
0.07 |
1.02 |
0.10 |
||
Sexual Abuse |
- |
- |
.17*** |
-6.71 |
.19*** |
-6.18 |
||
Emotional Abuse |
- |
- |
1.48** |
3.03 |
1.54*** |
3.19 |
||
Poverty |
- |
- |
.97 |
-0.26 |
.97 |
-0.23 |
||
Chaotic Home |
- |
- |
.89 |
-0.98 |
1.02 |
0.17 |
||
Gang Family |
- |
- |
1.35* |
1.95 |
1.06 |
0.33 |
||
Violent toward Family |
- |
- |
.75* |
-2.21 |
.74* |
-2.27 |
||
Delinquent Career |
||||||||
Age First Commitment |
- |
- |
- |
- |
.85*** |
-3.89 |
||
Prior Felony Adjudications |
- |
- |
- |
- |
1.53*** |
4.24 |
||
Prior Total Adjudications |
- |
- |
- |
- |
.65*** |
-5.74 |
||
Prior Out of Home Placements |
- |
- |
- |
- |
.96 |
-0.79 |
||
Gang Affiliation |
- |
- |
- |
- |
.76* |
-2.31 |
||
Instant Offense Gang-Related |
- |
- |
- |
- |
4.15*** |
10.37 |
||
Model χ2 |
71.45*** |
153.87*** |
337.9*** |
|||||
Log Likelihood |
-1434.68 |
-1393.47 |
-1301.46 |
|||||
Pseudo R2 |
.024 |
.052 |
.115 |
|||||
Note: OR=odds ratio, Z= z-score, ***p<.001, **p<.01, *p<.05 |
In model 1, African Americans were 157% more likely than non-Hispanic whites to be committed for a homicide. The effect declined to 114% in model 2 then jumped to 123% in model 3. In model 1, Hispanics were 158% more likely than non-Hispanic whites to be committed for a homicide. The effect declined to 119% in model 2 and 92% in model 3. Social risk factors evinced mixed relationships to homicide. Sexual abuse was associated with an 83% reduced likelihood in model 2 and 81% reduced likelihood in model 3. Emotional abuse was associated with a 48% increased likelihood in model 2 and 54% increased likelihood in model 3. Youth who were violent toward family members were 25% and 26% less likely to be committed for a homicide. Physical abuse, poverty, and chaotic home were not significantly associated with homicide perpetration.
In the full model (model 3), earlier commitment to a juvenile justice facility was associated with homicide by 15% per year. Thus those first committed at age 10 years were 60% more likely to commit a homicide than those first committed at age 14 years. Prior felony adjudications were associated with a 53% increased likelihood of homicide but prior total adjudications were associated with a 35% reduced likelihood. Youth who were gang affiliated were 24% less likely to be committed for a homicide; however, those whose offense was gang-related were 315% more likely to be committed for a homicide. Overall, the addition of social risk and delinquent career factors reduced the effect of African American status on homicide from 157% to 123% increased likelihood and for Hispanics from 158% to 92% increased likelihood.
Sociological explanations of homicide traditional cite socioeconomic and social disadvantage differences across racial and ethnic groups as the primary explanation for racial and ethnic differentials in homicide [for a review of diverse approaches, see 9]. Yet these social risk factors did not reduce the racial and ethnic effects from significance, and poverty was not associated with homicide in any model. Indeed, aside from those who perpetrated their crime expressly in the context of gang-related activity, none of the covariates were as strongly associated with homicide as African American and Hispanic status.
The current findings are fully consistent with other research of at-risk youth and their involvement in homicide offending and also homicide victimization. For instance, Farrington and colleagues [10] analyzed data from the Pittsburgh Youth Study and found that African American status was the second strongest predictor of homicide offending with an odds ratio of 4.6 and only was exceeded by selling hard drugs which increased the odds of homicide offending by an odds ratio of 5.5. Other factors including gun carrying, weapons use, gang membership, persistent drug use, and marijuana selling were also significantly linked to homicide offending albeit at lower risk than African American status (analyses were limited to blacks and whites). Additionally, Farrington and colleagues found that African American status was the strongest predictor of homicide victimization with an odds ratio of 11.1. Because crime is overwhelmingly intraracial, disproportionately high homicide offending among a specific racial or ethnic group will also disproportionately be associated with similarly high homicide victimization.
Of course, while the current study had a range of social risk and delinquent career controls, there are additional correlates of homicide offending that have been shown to mediate racial and ethnic differences in lethal violence. These include lifestyle factors, structural disadvantage, health behaviors, medical care, and others.
CONCLUSION
For a variety of historical and contemporary reasons, the association between race, ethnicity, and violent crime is controversial in the United States. At times, race differences in homicide offending and victimization are minimized due to political correctness, or, more commonly in academic criminology, attributed to socioeconomic and cultural disadvantage. The current brief report found that African Americans and Hispanics were strongly associated with being committed to a confinement facility for homicide, and the effects withstood strong confounds. It is hoped that the current report brings additional empirical clarification to discussions about race, ethnicity, and homicide particularly among clinical/correctional samples.