score) are entered simultaneously into a third multiple regression, the mediator (i.e., DTS–total score) must be significantly associated with the criterion (i.e., MMM–Coping motives), as depicted by Figure 1–Path B; and the relation between the predictor and criterion is either diminished (partial mediation) or eliminated (full mediation) as a function of the inclusion of the mediator, as depicted by Figure 1- Path C’. Marijuana use frequency (MSHQ–marijuana use in past 30 days) and non-criterion marijuana use motives (i.e., Enhancement, Conformity, Expansion, and Social motives) were entered as covariates in the regression equations when marijuana use coping motives was the criterion variable; and only marijuana use frequency was entered as a covariate when DT was a criterion variable.2 An alpha level of .01 was applied to all analyses to control for family-wise error attributable to the number of proposed analyses conducted (Tabachnick & Fidell, 2007).
