For significant findings, effect sizes were calculated using the “multilevel tools” (mlt) package in Stata 12.1 (Mohring & Schmidt, 2012), which computes effect sizes according to the Snijders &Bosker (1994) method. Between-subject effects (R22) are reported as the proportion of the between-subject variance accounted for by the effect. The effect size of Time (R21) is reported as the proportion of within-subject variance accounted for by Time.Significant Group moderator interaction effects werefollowedby 1) tests of Group differences at “low” (1 SD below the mean) and “high” (1 SD above the mean) values of the moderator variable, and2) tests of simple effects of the moderator within each Group. Sig- nificant interactions with Time were followed by tests of the simple effects of the moderatoron outcome ateach time point, followed by contrasts (Bonferroni-corrected for multiple comparisons) at the mean value of the moderator and ± 1 SD.