Second, we only focused on positive parenting because our measures of negative parenting were weak (only included a few items assessing strictness and spanking), only available for a subset of the sample, and lacked adequate reliability. However, parenting is a multidimensional construct, and positive and negative parenting may have differential effects on offspring outcomes (Dallaire et al., 2006). Thus, future studies should consider additional dimensions of parenting to further understand the effect of parental ADS on adolescent outcomes via parenting. In addition, although we used a latent variable approach to measure positive parenting that capitalized on multiple dimensions of positive parenting behaviors and helped account for measurement error, we acknowledge that some of the factor loadings were low (e.g., .57 for maternal involvement). This limitation regarding the measurement of positive parenting may have attenuated the effects observed in our analysis.
