Table 2.

Model fit indices for deciding on the number of classes of postpartum depressive symptoms: growth mixture models with linear (slope) and quadratic effects

Number of classes	BIC	LMR-LRT p	BLRT p	Entropy	n(%) of women in each class
1	29,156.61	–	–	–	1427 (100)
2	28,965.78	.02	<.001	.86	1318 (92.3), 109 (7.7)
3	28,904.68	.04	<.001	.89	1309 (91.7), 77 (5.4), 41 (2.9)
4	28,885.76	.16	<.001	0.88	1268 (88.8), 72 (5.1), 46 (3.2), 41 (2.9)

For each model, all classes had free (equal) intercept growth factor variances and the slope and quadratic growth factor variances fixed to zero. For the 1-, 2-, and 3-class models, the number of random starts was 100 and the number of final stage iterations was 20. For the 4-class model, the number of random starts was increased to 500. Values in italics indicate the final model

BIC Bayesian Information Criterion, LMR-LRT Lo-Mendell-Rubin Likelihood Ratio Test, BLRT Bootstrapped likelihood ratio test