TABLE 3.

Latent class solution of student behaviors at the year 2018.

Model	LL	npar	AIC	BIC	CAIC	AWE	BF (K, K+1)	cmP(K)	SIC	exp(SIC-max)	Entropy
1-class	−7907.24	6	15826.47	15852.51	15858.51	15908.55	0.000	0.000	−7926.3	2.43E-150	-
2-class	−7598.58	13	15223.15	15279.58	15292.58	15401.00	0.000	0.000	−7639.8	6.257E-26	0.611
3-class	−7528.60	20	15097.19	15184.00	15204.00	15370.80	0.000	0.000	−7592	3.562E-05	0.761
4-class	−7496.16	27	15046.32	15163.51	15190.51	15415.70	4098.013	1.000	−7581.8	1	0.760
5-class	−7482.29	34	15032.58	15180.15	15214.15	15497.72	67255.73	0.000	−7590.1	0.000244	0.668
6-class	−7471.21	41	15024.43	15202.38	15243.38	15585.34	0.000	0.000	−7601.2	3.628E-09	0.657

LL = loglikelihood; npar = number of free parameters in the model; AIC = Akaike information criterion; BIC = Bayesian information criterion; CAIC = Bozdogan’s consistent AIC; AWE = Approximate weight of evidence criterion (Masyn, 2013); BF = Bayes factor; cmP(k) = correct model probability; SIC = Schwartz information criterion. Entropy estimates are presented to realize latent class separation; this information did not contribute to the class enumeration process. Bolded estimates show preference by a criterion for a given class solution.