Operationalization and measurement of compulsivity across video gaming and gambling behavioral domains

BMC Psychology

Table 3 Goodness-of-fit indices for the different models

	χ²	df	CFI	TLI	RMSEA	SRMR	ΔCFI	ΔRMSEA
Full version (GRACC90)
Full sample
M1. 1 factor	11,112.9	3915	0.971	0.970	0.054	0.039
M2. 2 factors	8,752.9	3826	0.980	0.979	0.045	0.029
M3. 3 factors	7,019.7	3738	0.987	0.986	0.037	0.023
Subsamples
M4. Gambling sample	6,309.6	3915	0.984	0.984	0.044	0.033
M5. video games sample	7,504.3	3915	0.968	0.967	0.054	0.052
Invariance by type of activity
M6. Equal form	13,904.2	7830	0.977	0.976	0.050	0.044
M7. Equal loadings and thresholds	14,325.2	8188	0.977	0.977	0.049	0.044	0.000	–0.001
Short version (GRACC18)
Full sample
M8. 1 factor	501.0	135	0.996	0.996	0.066	0.013
Subsamples
M9. Gambling sample	298.0	135	0.997	0.996	0.062	0.014
M10. video games sample	407.6	135	0.994	0.993	0.080	0.019
Invariance by type of activity
M11. Equal form	713.6	270	0.995	0.995	0.072	0.017
M12. Equal loadings and thresholds	723.7	340	0.996	0.996	0.060	0.018	0.001	–0.012

Notes. df = degrees of freedom; CFI = comparative fit index; TLI = Tucker-Lewis index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual; Δ = increment in fit index with respect to previous model. All p-values for the chi-square test were < 0.001

ISSN: 2050-7283