In the original manuscript, two girls in the comparison group were erroneously coded as boys. Thus, of the 284 comparison youths included in the analyses, 229 were male and 55 were female. In analyses with the corrected sex variable, the patterns of findings were identical to those presented in the original manuscript with one exception: in the model for positively biased self-perceptions in the behavioral domain, the test of sex differences in within-time associations was now significant (p = .049). However, the change in the comparative fit index remained the same (.003) and was below the .01 threshold thought to reflect meaningful differences in model fit (Cheung & Rensvold, Reference Cheung and Rensvold2002). Follow-up analyses examining sex moderation of within-time associations at each time point provided evidence of invariance across sex at Time 1, scaled Δχ² (6) = 8.84, p = .18; Time 2, scaled Δχ² (6) = 10.90, p = .09; and Time 3, scaled Δχ² (6) = 2.67, p = .85; but not at Time 4, scaled Δχ² (6) = 13.82, p = .03. Follow-up nested model comparisons of individual within-time associations at Time 4 indicated that poor social skills were more strongly related to rejection among girls (β = –0.51, p < .001) than among boys (β = –0.39, p < .001), scaled Δχ² (1) = 5.00, p = .03, at this time point. All other within-time associations at Time 4 were invariant across sex. We regret this error and any problems it may have caused.