The MIMIC model without binge drinking showed the best fit to the data, however, the MIMIC model results were not appropriate to test our hypotheses based on factor analyses in our prior report (Beseler et al., 2010). Tests of these hypotheses required an analytic approach such as LCA, which could divide the sample into classes or groups for comparisons regarding drinking patterns--specifically frequency of binge drinking--and assessment of whether any differences in drinking patterns could be explained by personality traits and drinking motives.
