An advantage of the prospective design of this study was that it established temporal precedence between peer delinquency, and the outcomes of interest assessed 3 years later. Cohen and Cohen (1983) recommended guideline of using values corresponding to one standard deviation (SD) above and below the sample mean for peer delinquency was used to probe the interaction. Given that the value corresponding to one standard deviation below the mean was outside the range of possible responses, a value of −0.9 was used instead. Finally, an online computational tool was used to probe significant interactions as well as providing a formal analysis of the regions of significance (Preacher et al. 2006). This provides a significance test of the association between the moderator and the outcome at all values of the independent variable that fall within the range of interest.
