TABLE 4.

Fit indexes for determining optimal number of latent classes.

Class	K	Log(L)	AIC	BIC	aBIC	Entropy	LMR(P)	BLRT(P)	Conditional probabilities
C1	9	-20010.134	40038.27	40084.95	40056.36	N/A	N/A	N/A	N/A
C2	12	-19614.072	39252.15	39314.39	39276.27	0.98	0.03	<0.001	0.937/0.063
C3	15	-19329.220	38688.44	38766.24	38718.60	0.99	0.58	<0.001	0.059/0.010/0.931
C4	18	-19160.985	38357.97	38451.34	38394.16	0.98	0.24	<0.001	0.044/0.911/0.042/0.004
C5	21	-19064.019	38170.04	38278.96	38212.26	0.99	0.22	<0.001	0.005/0.906/0.032/0.053/0.003

K refers to a number of freely estimated parameters; Log (L) refers to Loglikelihood; LMR refers to Lo–Mendell–Rubin Likelihood Ratio Test; BLRT refers to Bootstrap Likelihood Ratio Test. The bold values represent those of the best fitting model.