Table 1.
Model fit indices derived from latent class analysis on models with 1–5 classes.
Model | k | AIC | BIC | aBIC | Entropy | LMR | BLRT | Class probability |
---|---|---|---|---|---|---|---|---|
1 | 10 | 7,978.61 | 8,027.41 | 7,995.65 | ||||
2 | 21 | 5,513.19 | 5,615.68 | 5,548.99 | 0.95 | < 0.001 | < 0.001 | 208/765 (0.21/0.79) |
3 | 32 | 5,205.72 | 5,361.90 | 5,260.26 | 0.92 | < 0.001 | < 0.001 | 67/200/706 (0.07/0.20/0.73) |
4 | 43 | 5,186.05 | 5,395.91 | 5,259.34 | 0.89 | 0.23 | < 0.001 | 164/700/44/65 (0.17/0.72/0.05/0.07) |
5 | 54 | 5,173.52 | 5,437.06 | 5,265.56 | 0.90 | 0.51 | < 0.001 | 30/152/48/703/40 (0.03/0.16/0.05/0.72/0.04) |
AIC, Akaike information criterion; BIC, Bayesian information criterion; aBIC. sample size adjusted BIC; LMR, Vuong-Lo-Mendell-Rubin likelihood ratio test; BLRT, bootstrapped likelihood ratio test. The three latent classes were regarded as the best model, highlighted in bold.