Ordinary least squares (OLS) regression was run using the regression procedure in SAS 9.1.3 (SAS Institute Inc., 2004). Intentions to use cigarettes and intentions to use alcohol were analyzed in separate models. Of interest were two two-way interactions (agentic goals × perceived peer attitudes and communal goals×perceived peer attitudes)1. In cases where interactions contained continuous moderators (e.g., social goals), Cohen and Cohen’s (1983) recommended guideline of using values corresponding to one standard deviation above and below the sample mean was used to probe the interactions. We probed marginally significant interaction terms (p ≤ .10) because statistical interactions are often difficult to detect in the social sciences (McClelland & Judd, 1993) and because we had a priori predictions about the nature of the interactions. Squared semi-partial correlations (sr2) are presented to provide information about effect sizes. In addition, given that some research suggests gender differences in social goals (Locke, 2003; Salmivalli, Ojanen, Haanpaa, & Peets, 2005) and in substance use (e.g., Luthar & D’Avanzo, 1999), gender and gender interactions were considered. None of the first order effects and interactions with gender were statistically significant for either the cigarette or the alcohol models, and accordingly, gender was not considered further.
