TABLE 4.

Goodness-of-fit indices and model comparisons for tested models.

	WLSMV χ2	df	p	CFI	TLI	RMSEA (90% CI)	BIC
Model A	2071.315	189	<0.001	0.944	0.938	0.065 (0.063, 0.068)	92,348.487
Model B	1804.539	188	<0.001	0.952	0.946	0.060 (0.058, 0.063)	92,144.291
Model C	1809.732	188	<0.001	0.952	0.946	0.061 (0.058, 0.063)	92,123.284
Model D	1670.857	186	<0.001	0.956	0.950	0.058 (0.056, 0.061)	92,016.280
Model E	1051.015	186	<0.001	0.974	0.971	0.044 (0.042, 0.047)	91,596.218
Model F	1414.820	176	<0.001	0.963	0.956	0.055 (0.052, 0.057)	91,782.712
Model G	738.317	168	<0.001	0.983	0.979	0.038 (0.035, 0.041)	91,407.125
Model I	745.651	173	<0.001	0.983	0.979	0.037 (0.035, 0.040)	91,370.907

WLSMV, weighted least squares with mean and variance adjustment; df, degree of freedom; TLI, Tucker–Lewis index; CFI, comparative fit index; BIC, Bayesian information criterion; RMSEA, root mean square error of approximation; CI, confidence interval; Model A, single-factor model; Model B, the two-factor model in the Beck et al. (1996) study; Model C, the two-factor model in Huang and Chen (2015); Model D, the three-factor model in Wu’s (2010); Model E, the three-factor model in the Zhu et al. (2018); Model F, the bifactor model in Ward’s (2006); Model G, a bifactor (S.I-1) model in the Faro and Pereira’s (2020); Model I, the bifactor model initially identified in the current study.