main effect, but also for the extent to which it modified relationships between heavy drinking and alcohol outcomes. We then tested for significant differences in the coefficient representing the race/ethnicity effect on the intercept in models with and without the candidate mediator of interest. Statistical testing was performed using bootstrap methods and randomly re-sampling 1000 replications. Those variables shown to reduce the coefficient representing the effect for race/ethnicity were retained for the final analysis (below). We used a piecemeal approach informed by Baron and Kenny (1986) for preliminary tests because this approach was best suited to capturing our specific hypotheses (Muller et al., 2005) and because available macros for testing mediated moderation are not yet suitable for multicategory mediators (Hayes, 2014). These models do however essentially test mediated moderation (Muller et al., 2005): that is, how socioeconomic status, prejudice, unfair treatment, norms, age, and marital status may explain the interaction effect for race/ethnicity and alcohol use in predicting alcohol problems.
