After adjusting for sex and age at initiation, the MIMIC model showed that a single factor best fit the data after excluding binge drinking (Table 2). Although sex was not a statistically significant covariate in the model, the model fit the data marginally better when it was included. After adding binge drinking, drinking more than intended and binge drinking loaded separately on a single factor, but the correlation between the two factors was very high (r = 0.90). Age at initiation was significantly associated with both factors and the odds of binge drinking were higher for males than for females (OR = 1.96, 95% confidence interval 1.06-3.62). Nonetheless, the best fitting MIMIC model did not include binge drinking as a criterion after controlling for sex and age at initiation. LCA results showed that a 3-class solution without binge drinking significantly fit the data better than a 2-class solution (Table 2). Furthermore, a 2-class solution without binge drinking fit the data significantly better than a 2- or 3-class model that included binge drinking (Table 2). A 4-class solution identified a class
