Table 2.
Model selection and fit indices
| Classes | Log likelihood | Free parameters | AIC | BIC | BIC (Sample-Size Adjusted) | Entropy | LMR p value for (k-1) classes |
|---|---|---|---|---|---|---|---|
| 1 | −24,695.3 | 13 | 49,416.61 | 49,494.25 | 49,452.95 | a | |
| 2 | −24,234.91 | 17 | 48,503.82 | 48,605.36 | 48,551.35 | 0.913 | 0.005 |
| 3 | −24,012.61 | 21 | 48,067.23 | 48,192.66 | 48,125.94 | 0.924 | 0.049 |
| 4 | −23,845.44 | 25 | 47,740.89 | 47,890.21 | 47,810.78 | 0.892 | 0.236 |
Model with the best fit is shown in bold
AIC Akaike information criteria, BIC Bayesian information criteria, LMR p value, Lo, Mendell, and Rubin likelihood ratio test p value
aNot estimable for a 1-class model