challenge and SRE LR scores predicting the usual drinks per occasion during the six-month period prior to follow-up (or Time 2 or T2). These were followed by five regressions using only SHAS and SRE LR as predictors. The first regression entered only SRE LR; the second only SHAS; the third entered both SHAS and SRE LR simultaneously; the fourth entered SHAS first followed by SRE LR; and the fifth entered SRE LR first followed by SHAS. These regressions allowed for the computation of a set of R2’s that represented the unique variance of T2 drinking predicted by SRE LR when SHAS was also in the equation; the unique variance of T2 drinking predicted by SHAS when SRE LR was also in the equation; and the variance of T2 drinking that was mutually predicted by both SRE LR and SHAS. Multilevel modeling was used to evaluate the role of sibships on the results (Muthén and Satorra, 1995).
