Table 1.
Model Fit of the Latent Profile Analyses With Up to Six Latent Classes (N = 215).
| Number of Latent Classes | ||||||
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | |
| Log-Likelihood | −1,880.57 | 1,796.03 | −1,759.05 | −1,738.43 | −1,720.87 | −1,707.67 |
| No. of Free Parameters | 8 | 13 | 18 | 23 | 28 | 33 |
| BIC a | 3,804.68 | 3,661.88 | 3,614.77 | 3,600.39 | 3,592.12 | 3,592.57 |
| Adjusted BIC | 3,779.33 | 3,620.68 | 3,557.73 | 3,527.51 | 3,503.29 | 3,488.00 |
| AIC | 3,777.71 | 3,618.06 | 3,554.09 | 3,522.87 | 3,497.74 | 3,481.34 |
| (−2)*Log-Likelihood Difference b | – | 169.06 | 73.96 | 41.24 | 35.12 | 26.40 |
| LMR LRT, p Value c, e | – | <.001 | .008 | .263 | .388 | .034 |
| Bootstrap LRT, p Value c, d | – | <.001 | <.001 | <.001 | <.001 | <.001 |
| 1 – Entropy | – | .873 | .828 | .826 | .844 | .827 |
AIC, Akaike’s information criterion; BIC, Bayesian information criterion; LRT, likelihood ratio test. aIncremental changes of BIC < 2 are considered marginal (Kass and Raftery (43), p. 777). bDifference between models with (k − 1) and k classes. cLMR, Likelihood ratio test according to Lo, Mendell, and Rubin (38). dLRT according to Nylund et al. (39). eIf < .05, a model with k latent classes will fit significantly better than a model with (k − 1) latent classes.