. Author manuscript; available in PMC: 2020 Nov 1.

Published in final edited form as: Psychol Addict Behav. 2019 Aug 26;33(7):603–615. doi: 10.1037/adb0000505

Table 3.

Model Fit Indices

	X² (df), p-value	X² Diff. Test	RMSEA	CFI	SRMR	AIC	BIC

Model 1	167.80 (48,)<.001	-	.08 (.07 & .09)	.95	.08	36507.99	37092.14
Model 2	179.22 (56), < .001	8.63, > .05	.08 (.06 & .09)	.95	.08	36525.77	37078.129
Model 3	197.79 (75), <.001	19.09, > .05	.07 (.05 & .08)	.95	.08	36513.89	36990.75
Model 4	272.00 (111), <.001	78.22, < .05	.06 (.05& .07)	.93	.09	36507.79	36841.59
Model 5	320.08 (126), <.001	48.39, < .05	.06 (.05 & .07)	.91	.10	36549.44	36823.63
Model 6	379.61 (138), <.001	58.58, < .05	.07 (.06 & .08)	.89	.15	36608.03	36834.54

Note. Due to the use of MLR estimation in analyses, Chi-square difference tests accounted for scaling correction. Chi-square difference test results for each model were compared to the previous model in the Table. Significant Chi-square indicates significantly worse fit for the more constrained model. Bold text indicates the strongest fit for each index. Model 1 = All relationships (autoregressive, cross-lagged, and covariances) unconstrained; Model 2 = autoregressive effects constrained; Model 3 = autoregressive and cross-lagged effects constrained; Model 4 = autoregressive, cross-lagged, and covariate effects constrained; Model 5 = autoregressive, cross-lagged, covariate, and across-time covariance effects constrained; Model 6 = autoregressive, cross-lagged, covariate, across-time covariance, and within time covariance effects constrained. Df = degrees of freedom; RMSEA = Root Mean Square Error of Approximation; CFI = Comparative Fit Index; SRMR = Standardized Root Mean Squared Residual; AIC =Akaike Information Criterion; BIC = Bayesian Information Criterion.