Table 2.

Gene–environment interactions between DRD4 and positive as well as a negative peer social preference in predicting conduct problems and prosocial behavior: fit statistics and nested model comparisons

Model	χ 2	df	CFI	SRMR	Comp.	Δχ 2	Δdf	p
Positive social preference
Total sample
1. Base model	91.14	33	.95	.05
2. Time constraints	111.67	47	.94	.05	1 vs. 2	20.37	14	0.119
DRD4-7r versus DRD4-no7
3. No constraints	172.69	94	.94	.07
4. Non-hypothesized paths equal	183.71	103	.94	.08	3 vs. 4	9.87	9	0.361
5. GxE: positive social preference → prosocial behavior equal	183.31	104	.94	.08	4 vs. 5	0.06	1	0.805
6. GxE: positive social preference → conduct problems equal	186.87	105	.94	.08	5 vs. 6	5.70	1	0.017
Negative social preference
Total sample
1. Base model	84.66	33	.95	.06
2. Time constraints	96.27	47	.95	.05	1 vs. 2	14.84	14	0.389
DRD4-7r versus DRD4-no7
3. No constraints	184.49	94	.93	.07
4. Non-hypothesized paths equal	194.34	103	.93	.07	3 vs. 4	9.27	9	0.413
5. GxE: negative social preference → prosocial behavior equal	196.67	104	.92	.07	4 vs. 5	2.85	1	0.091
6. GxE: negative social preference → conduct problems equal	204.01	105	.92	.07	5 vs. 6	9.81	1	0.002

Δχ ² statistics are based on the Satorra–Bentler Chi square difference test