To test for potential differences in the associations between parental ADS, parenting behaviors, and adolescent outcomes across ethnicity and adolescent gender, we first conducted multigroup confirmatory factor analyses (MGCFA) to consider measurement invariance for the paternal and maternal positive parenting latent factors across the comparison groups (i.e., EA vs AA; Male vs Female), followed by multigroup SEM analysis with Wald chi-square test of parameter equalities. MGCFA compared two alternative models, one with factor loadings for the latent constructs freely estimated across groups and another model where factor loadings were constrained to equality across groups. To the extent that the two alternative models do not differ significantly in model fit, as indicated by a non-significant chi-square difference test and/or difference in Comparative Fit Index (ΔCFI) smaller than .01 in absolute value (Cheung & Rensvold 2002), there is evidence of metric measurement invariance, which is sufficient for group comparisons in structural path coefficients (Byrne & Watkins, 2003; Steenkamp & Baumgartner, 1998). To test for potential differences in path coefficients between males and females, multigroup analysis was conducted by removing gender from the
