The social-cognitive information-processing model explains decision-making processes and cognitive structures children use to respond to social conflict (Huesmann, 1998). The present study examined links between two steps in this model: the script search and retrieval step and the script evaluation step. In the first hypothesis, I predicted that fantasizing about aggression, believing aggressive responses will make repeated provocation stop, and approving of retaliation would predict the frequency of self- and teacher-reported actual direct aggression. In the second hypothesis, I predicted that the relations in the first hypothesis would be mediated by aggressive response selection, which was measured by the frequency with which children chose aggressive responses to hypothetical repeated provocation scenarios. The first hypothesis was supported and replicated previous research. The second hypothesis was not supported.
A supplementary hypothesis was developed to explore the possibility that mediation was not observed because of differences in the level of measurement of the constructs. Crick and Dodge (1994) advanced a distinction between global-level social knowledge, and online processing actions, which are the ways in which children think and make decisions in real time about specific social situations in which they are currently involved. The mediator in the second hypothesis, aggressive response selection, may have measured aggression-supporting online processing actions that were employed only by children who had aggression-supporting social cognitions because children who lack aggression-supporting social cognitions may have filtered out aggressive responses upstream in their social-cognitive information processing. This supplementary hypothesis was not supported.
Because the social-cognitive information-processing model posits sequential relations between aggressive fantasy, outcome expectancy, and aggressive response selection, I developed another supplementary hypothesis to test for mediation of the relations between aggressive fantasy, and aggressive response selection through outcome expectancy. This supplementary hypothesis was supported for boys. Higher levels of aggressive fantasy were related to higher levels of outcome expectancy, which were in turn related to higher levels of aggressive response selection. Limitations and future research needs are discussed.