After testing the measurement models for emotion regulation and aggressive behavior, the first mediation analysis proceeded as follows: (1) Time 1 peer victimization was examined as a predictor of aggressive behavior at Time 3, controlling for Time 1 aggressive behavior; (2) Time 1 peer victimization was examined as a predictor of Time 2 emotion dysregulation, controlling for Time 1 emotion dysregulation; (3) Emotion dysregulation at Time 2 was evaluated as a predictor of aggressive behavior at Time 3, controlling for Time 1 aggressive behavior; (4) the full mediation model was examined to evaluate the hypothesis that emotion dysregulation mediates the relationship between peer victimization and aggressive behavior. Sobel’s standard error was used to test the significance of the intervening variable effect (Sobel 1982). The product of coefficients approach is associated with low bias and Type 1 error rate, accurate standard errors, and adequate power to detect small effects (MacKinnon et al. 2002). A separate set of mediation analyses were conducted using the same approach outlined above to evaluate whether emotion regulation mediated the association between stressful life events at Time 1 and aggressive behavior at Time 3, controlling for Time 1 aggression.
