Binge drinking did not improve the model fit either in the MIMIC or LCA, and this may be due to the fact that the two most severe classes did not differ by endorsement of weekly binge drinking. Our previous report suggested the presence of a second dimension after adding binge drinking as a criterion to the AUD criteria in a factor analysis, although the second factor was highly correlated with the first (Beseler et al., 2010). The MIMIC analysis suggested measurement invariance or population heterogeneity due to age of initiation. Early age at initiation has been associated with behavioral undercontrol/disinhibition longitudinally (Iacono et al., 2008). The second factor uncovered in our prior report could possibly be explained by a clear separation between problem and non-problem drinkers. As no overlap occurred between class 1 and class 3, this finding argues against binge drinking falling on a continuum of severity with other criteria. This result supports previous findings of non-dimensionality of binge drinking, which may serve better as a risk factor for an AUD in the presence of other risk factors (e.g., personality traits and drinking motives), and may not add linearly to the DSM-IV criteria as a severity indicator.
