Table 4.

Results of factorial invariance testing for the dysphoria model by sex.

Type of invariance	χ2	df	CFI	TLI	RMSEA [90% CI]	χ2 difference testa	df	p
Configural	364.903**	276	.986	.986	.040 [.028, .051]	---	---	---
Weak factorial/matric	355.216**	289	.989	.990	.034 [.020, .045]	10.332	13	.667
Strong factorial/scalar	371.943**	302	.989	.990	.034 [.020, .045]	14.679	17	.619
Strict factorialb	371.943**	302	.989	.990	.034 [.020, .045]	22.517	17	.166
Factor variance	362.178*	306	.991	.992	.030 [.014, .042]	3.539	4	.472
Factor covariance	340.518	312	.995	.996	.021 [.000, .035]	4.730	6	.579
Factor mean	335.966	316	.997	.997	.018 [.000, .033]	3.089	4	.543

Note. df = Degrees of freedom; CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; RMSEA = root mean square error of approximation;

CI = confidence interval.

^**p < .01,

^*p < .05

^aχ² difference test computed using the DIFFTEST option in Mplus.

^bFit statistics for the strict factorial invariance model are the same as the strong factorial/scalar invariance model but the model was compared to one that allowed residual variances to be freely estimated across groups, as indicated with results of the χ² difference test.