The results in Figure 3 regarding LR are similar to those reported in earlier iterations of the COGA Youth Panel and in other populations (e.g., Schuckit et al., 2004, 2011; Schuckit and Smith, 2006). One exception is that while studies of adults generally show a direct path from LR to COPE, SEMs using younger subjects rarely show that direct link (e.g., Schuckit et al., 2009a, 2010, 2012). Perhaps the relationship of LR to COPE becomes stronger as drinkers get older and have had more experience with alcohol, or the differences might reflect age-related increases in social desirability biases where older heavy drinkers see drinking to cope as a more acceptable explanation for their drinking behaviors. Another result worthy of comment in Figure 3 is the absence of a consistently significant relationship of IMPULSE to PEER, even when adolescent and adult Barratt scales were evaluated separately or when Figure 3 was reevaluated using a latent construct for IMPULSE generated from the three subscales of that measure.
