Finally, to evaluate Hypothesis 4 outside the context of the SEM, we determined the relative performance of using Baseline SRE-5-LR alone to predict ALC PROBS, Baseline impulsivity alone as a predictor, and using both predictors together. In step 1 of a hierarchical regression analysis SRE5-LR was entered, producing an R2=.083 (p<.001, f2=.09) in predicting ALC PROBS; and in step 2, after adding impulsivity the change in R2 =.04 (p<.001, f2=.04). Subsequently, in a separate hierarchical regression analysis, in step 1 impulsivity was entered first and yielded an R2 =.056 (p<.001, f2=.06) in predicting ALCPROBS; then in step 2 SRE5-LR was entered and the change in R2=.067 (p<.001, f2=.07). When both SRE5-LR and impulsivity were entered simultaneously R2=.122 (p<.001, f2=.139). The results indicate that SRE5-LR and impulsivity contributed independently to alcohol problems.
