A four-class solution provided the most parsimonious, informative, and readily interpreted fit to the data (see Table 2 for model fit statistics). Although indicators of good model fit (e.g. AIC, BIC) decreased as more classes were included, for solutions with more than four classes, some of the class prevalences were very low (two classes ≤ 5% in the five- and six-class solutions), precluding meaningful interpretation. Therefore a four-class solution was chosen. Fig. 1 provides a graphic overview of the LCA results. Class 1, or the “lowest severity” class, was the largest at (n = 1134, 44% of sample) and included low to moderate endorsement on all alcohol and PTS symptoms (cutoff: endorsement probability [EP] < 0.60). Across all classes, the most commonly endorsed alcohol symptoms were those reflective of symptom 1 (alcohol is taken in larger quantities and over a longer period of time than intended), symptom 2 (more than once tried to cut back on drinking but could not), symptom 8 (more than once got into situations after drinking that increased chances of getting hurt) and 10 (had to
