For hypothesis 1, we assessed mediation with the total indirect effect (product of coefficients) approach (MacKinnon, 2000) and computed 95% Monte Carlo confidence intervals to test the indirect effects (Mackinnon et al., 2004; Preacher and Selig, 2012). Conceptually, a total indirect effect represents the amount of the association between two variables that is explained by mediating variables. In multiple-mediator analysis, the total indirect effect is the sum of the specific indirect effects involving each mediator. SEM allowed for the simultaneous computation of all indirect effects for this study (i.e., for the multiple mediators, outcomes, and groups) with a single statistical model.
