Practice Profile: Targeted School-Based Social Information-Processing Interventions for Aggression

Wilson and Lipsey (2006) reviewed studies that met the following definitional criteria of social information-processing:

1. The program includes training in at least one of the social information-processing steps: 1) encoding situational and internal cues, 2) interpreting the situational and internal cues, 3) choosing or clarifying a goal, 4) producing or accessing possible responses to meet the goal chosen, 5) selecting a response, and 6) executing the behavior. 
2. Rather than targeting specific behavioral skills, the program emphasizes cognitive–thinking skills. Through teaching cognitive and thinking skills, social information-processing interventions aim to prove a student’s ability to process information in a variety of social situations.
3. The program includes structured tasks and activities to teach cognitive skills to students. Students then have the opportunity to apply the skills learned to actual social situations. 

If an intervention met the definitional criteria set, then the social information-processing component had to be the focus of the intervention. These interventions had to be delivered to school-aged children during regular school hours. Special education classrooms and alternative schools were eligible school settings, whereas classrooms in residential facilities and for afterschool programs were not eligible. Further, the intervention had to be delivered to students who were selected for treatment as a result of a risk factor linked to antisocial behavior or to students who were displaying problem behaviors. The evaluation had to measure at least one outcome variable of aggressive behavior, which could include violence, aggression, fighting, person crimes, disruptive behavior problems, acting out, conduct disorder, and/or externalizing problems. Finally, only studies with a comparison group were eligible. Both experimental and quasi-experimental designs were eligible; however, quasi-experimental designs had to match students or provide evidence of group equivalence on key demographic variables. Control groups could include placebo, wait-list, no treatment, or treatment as usual conditions.  

A comprehensive search of bibliographic databases, bibliographies of previous meta-analyses, and literature reviews was conducted. Both published and unpublished reports were included in the search, which was not limited to the English language. The search strategy used three categories of search terms to index articles for each specific database. These categories included a population of interest, risks or outcomes, and treatment/evaluations.  

The search yielded 68 eligible reports, which gave the results of 47 unique research studies. The majority of the studies used random assignment of individual students. Programs were mainly conducted using group formats in regular schools; however, seven programs were either one-on-one or included individual as well as group sessions. Five programs were conducted with special education students. Treatment duration ranged from 8 weeks to programs lasting the entire school year or more. Students’ ages ranged from 6 to 16, and the samples were predominately male. Of the samples who reported race/ethnicity, more than one half comprised minority youth: Black (32 percent), white (28 percent), Hispanic (4 percent), Asian (2 percent), and mixed (4 percent). The race/ethnicity of the remaining 30 percent of the samples was non-distinguishable. 

A random effects mean was used to examine the impact of selected and indicated social information-processing programs on aggressive and disruptive behavior. Thirty-two studies generated more than one effect size, whereas 15 studies provided only a single effect size. The most common type of measure was the teacher-report questionnaire, though many studies provided results for more than one aggressive behavior outcome. The multiple measures of aggressive and disruptive behavior within studies were either different types of aggressive behavior or were the same type of aggressive behavior reported by different informants. When studies reported results on different types of aggressive behavior, the effect size that closely resembled interpersonal physical aggression was selected. Each informant was used as a variable in the analysis, so when there was more than one effect size from the same informant, their mean value was used. In studies with multiple effect sizes reported by different informants, one effect size was selected by the informant who was more frequently represented in the data. This usually happened to be teacher reports, followed by self-report from participants. This procedure resulted in 47 effect sizes that are included in this meta-analysis.