The regression analyses in Tables 2 and 3 agree with other reports regarding the ability of a low LR to predict alcohol outcomes independently of the drinking practices at the time that alcohol challenges or SRE evaluations were determined (Schuckit et al., 2007, submitted; Trim et al., submitted). Thus, LR contributed additional predictive power beyond that expected from autocorrelations of the same behavior over time (e.g., the relationship between T1 and T2 drinking) (Malone et al., 2004).
