Comparisons of model fit were evaluated for linear, quadratic, and linear piecewise growth models, separately for males and females (see Supplemental Table 1). Similar to previously published findings using the same data (i.e., Dick et al., 2013), the piecewise models were marginally superior to the linear model and quadratic models in males and females. The piecewise growth model was selected for subsequent analyses for a more straightforward interpretation of covariate effects on heavy episodic drinking trajectories across two separate development epochs: (1) adolescence (age 15.5) to early adulthood (age 21.5) and (2) older adult (age 23.5 to age 27.5).
