We conducted follow-up analyses to check the robustness of our results. As noted above, there was a relatively large amount of missing data for parental assessments of ADS in our sample. We conducted additional analysis with the subsample (n = 706) of adolescents having complete data on both parents’ ADS, in order to examine whether the patterns of association are consistent or biased due to missing data. The overall pattern of associations was similar to results from analysis using the imputed datasets for the whole sample. Given that family processes may vary as a function of family structure/living arrangement, we also ran analysis with the subsample (n = 929) of families where adolescents lived with both biological parents to check whether the patterns of association were consistent. The overall pattern of associations was similar to those from the whole sample. Results from these follow-up analyses are not presented but available upon request.
