With regard to the effect sizes for the statistically significant mediated relationships, ISS mediated approximately 41% and 30% of the total relationship between PAS and past-year AUD for labeled and unlabeled individuals, respectively. Preacher and Kelley (2011) note that these ratios should be interpreted within the context of the size of the total effects. That is, larger mediation ratios are more “interesting” when accompanied by larger, as compared to smaller, total effects. For example, for past-year internalizing disorder the mediation ratio for ISS was larger for unlabeled individuals (82%) than labeled individuals (55%), yet the total effect for unlabeled individuals was quite small.
