Appendix A.
Number of Classes | Parameters | Log-Likelihood | AIC | BIC | LMR-LRT | p-value | Smallest Class Size |
---|---|---|---|---|---|---|---|
1 | 25 | −234,417 | 468,883 | 468,981 | - | - | - |
2 | 46 | −230,312 | 460,717 | 460,897 | 8,166 | < 0.01 | 3,571 (0.40) |
3 | 67 | −229,496 | 459,126 | 459,388 | 1,624 | < 0.01 | 1,732 (0.19) |
4 | 88 | −228,906 | 457,988 | 458,334 | 1,173 | < 0.01 | 1,690 (0.19) |
5 | 109 | −228,362 | 456,943 | 457,370 | 1,082 | < 0.01 | 160 (0.02) |
6 | 130 | −228,008 | 456,275 | 456,786 | 705 | < 0.01 | 162 (0.02) |
Note: Data are unweighted. Akaike Information Criterion (AIC), sample size adjusted Bayesian Information Criterion (BIC; Schwarz 1978; Sclove 1987), sample size adjusted Lo-Mendell-Rubin adjusted likelihood ratio test (LMR-LRT; Lo et al. 2001), and parametric bootstrap likelihood ratio test (parametric bootstrap LRT; Nylund et al. 2007). In the far right column, numbers represent number of respondents in the smallest latent class, with the percentages in parentheses. Bolded model indicates best fitting model.