We ran a series of multilevel linear mixed models (MLM) to examine the associations between problematic alcohol use polygenic scores (PGSPAU), parental separation, parental relationship discord, and the longitudinal course of HED. Our decision to use a multilevel approach over other longitudinal approaches such as latent growth curves was guided by the fact that an MLM can accommodate unequal spacing between time intervals and unbalanced data (ie, it does not require each individual to be observed at every chronological age).37
