An interesting study on race and aggression: Hawkins et al. 1991

This is one of the more interesting studies done on race differences in aggression, particularly due to its experimental nature. I felt like this study is deserving of a commentary blog post.

HAWKINS, J. D., VON CLEVE, E., & CATALANO, R. F. (1991). Reducing Early Childhood Aggression: Results of a Primary Prevention Program. Journal of the American Academy of Child & Adolescent Psychiatry, 30(2), 208–217. doi:10.1097/00004583-199103000-00008

For one, this study had a really decent sample for an experimental study. They acquired data from eight Seattle public elementary schools. The study was primarily on childhood aggression (which is strongly predictive of adulthood aggression) so they used first graders in the study. They split the sample into a full control school and a full experimental school. The experimental group received a number of special treatments including: parental training to help reduce family tension and problems at home, and teacher training to use more proactive classroom management, communicate better with students, and general cognitive social skills training.

The total sample was n = 458 students. They split the groups up by race, sex, and socioeconomic status. Experimental and control group were matched in demographic variables. The experimental group did tend to be more prosocial and have better family communication, supervision, and attachment to school, but were matched on 20/28 self-report variables.

The results are pretty simple: the treatment program had little to no effect on blacks and a significant effect on whites in most regards. Some interesting stuff here for the effect on males:

To determine the possible differential effects of the intervention on black and white boys, black male members of the experimental (N = 39) and the control groups (N = 31) were compared as were white male members of the experimental (N = 59) and control groups (N = 40) (Table 3). No significant differences on the CBCL scales were found for blacks, whereas white male control group subjects were scored significantly higher on the Aggressive scale ([El x = 6.88, [C] x = 13.4) and the global externalized deviance scale ([El x = 13.18, [C] x = 21.64), than were experimental group subjects at posttest. The project interventions resulted in significantly lower aggressiveness among white experimental group boys when compared with the control group. The interventions did not have significant effects on the aggressiveness or the externalized deviance of black boys.
To further investigate the issue of CBCL difference by race, black and white males were then compared within project conditions (Table 4). Within the experimental group, blacks scored significantly higher than whites on only two scales: Aggressive ([Whitelx = 7.72, [Blacklx = 15.55), and External ([White] x = 15.02, [Black] 52 = 25.47). Within the control group, there were no significant differences between blacks and whites on any of the 10 scales at posttest. Where significant race differences were found on the CBCL scales, black males were rated more negatively by their teachers than were white males. However, this appears to reflect significantly lower rates of teacher rated aggressiveness and external deviance among the white experimental group boys. Experimental and control group black boys were rated as having similar levels of aggressive and externalized deviant behaviors by their teachers.

So, that reduces one possible question I would have beforehand: did the groups already differ largely in aggression? Since the control groups for blacks and whites were not statistically significantly different on aggression scales, this gives some solid credence towards a theory of immutability of racial gaps in aggression. They later state,

As shown in Table 6, there were no significant differences between the experimental and control groups for black females. White female subjects of the control group were scored significantly higher on the Self-Destructive scale than were the subjects in the experimental group ([El x = 0.18, [C] = 0.86). Additionally, white female subjects in the control group were rated worse than white female subjects in the experimental group on two scales where difference? approached significance: Depressive ([El x = 1.43, [C]_X = 2.78) and Nervous-Overactive ([El x = 1.02, [C] X = 1.99). Again it appears that the project intervention had a beneficial effect for whites but not for blacks. As shown in Table 7, when black and white females were compared by project status, the control group white females were rated more anxious than were black females in the control group ([whitelx = 3.43, [blacklx = 1.24) (Table 7). Within the experimental group, blacks were rated significantly more negatively than whiteszn four of 10 scales: Depressive ([white] x = 1.49, [black] X = 2.79); NervousOveractive ([white] x = 1.13, [black] x = 2.03); Aggressive ([white] x = 3.77, [black] x = 11.02); and External ([whitelg = 8.91, [blacklx = 16.75). Findings for females were generally consistent with those for males in that, where significant differences were found, blacks were rated more negatively than whites, with the single exception of control teachers’ ratings of female anxiety.

Same thing as before but for females instead of males. Here is a table I made of the aggression results from the study. For some reason, they didn’t disaggregate the mean comparison results by race for females:

In the discussion of the paper they very eloquently dismiss a lot of the possible criticisms such as lack of pre-experiment testing (through a control group) and possible Hawthorne effects. They go on,

The results from this study provide limited support for the hypothesis that a combination of improved teacher instructional methods and training for parents in family management and relationship-building skills can reduce early antisocial behavior. However, the lack of significant intervention effects on the experimental group teacher ratings of black subjects is a cause for concern.

They seem to attribute this to teacher bias, which I would generally doubt. This study does show, however, that when blacks and whites were given an experimental treatment to reduce aggression, the program only affected whites. In consequence, the black white gap in aggression widened. The effects: an 86.6 percent increase in the race-aggression gap.

Some other studies on the topic:

Marcus, R. (2007). Aggression and Violence in Adolescence. Cambridge: Cambridge University Press. doi:10.1017/CBO9780511611292

Table from an earlier post using this review:

Mayberry, Megan & Espelage, Dorothy. (2007). Associations Among Empathy, Social Competence, & Reactive/Proactive Aggression Subtypes. Journal of Youth and Adolescence. 36. 787-798. 10.1007/s10964-006-9113-y.

Findings: “the aggression differences between blacks and whites are larger in reactive aggression (the hostile component).” Unfortunately authors don’t give group proportions, so it is difficult to find the SD difference properly. Disaggregated by sex, the SDs are pretty similar. For males, the SD difference by group was ~0.42 SDs; for females, the SD difference was ~0.75 SDs. The race and sex relations to reactive aggression were statistically significant in the ANOVA.

Wells, E. A., Morrison, D. M., Gillmore, M. R., Catalano, R. F., Iritani, B., & Hawkins, J. D. (1992). Race Differences in Antisocial Behaviors and Attitudes and Early Initiation of Substance Use. Journal of Drug Education, 22(2), 115–130. doi:10.2190/3bhh-3nat-bynk-d3vc 

Findings: “black males were nearly twice as aggressive as white males and that black females were nearly four times (!) as aggressive as white females. This is one of the few studies which also reported Asian American results as well – as we may expect, Asians were far less aggressive than both of the other groups. This is all after controlling for socioeconomic status.” (my own words.)

So, the evidence seems consistent in finding black white differences in aggression over time and the differences expand when given a large aggression treatment program.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s