Table 2.

Results from testing indirect effects via putative mediators using single-group design

Model	Indirect path	Indirect Effect (S.E)	Lower 2.5%	Upper 2.5%	p-value
1	Path a × b	.02 (.01)	.01	.05	.02*
	Path a × d	.01 (.01)	.00	.04	.22
	Path a × b × c	.01 (.00)	.00	.02	.10
2	Path a × b	.02 (.01)	.01	.05	.05*
	Path a × d	.02 (.01)	.00	.05	.09
	Path a × b × c	.01 (.00)	.00	.02	.10
3	Path a × b	.05 (.02)	.02	.10	.01*
	Path a × d	.05 (.02)	.01	.10	.03*
	Path a × b × c	.01 (.01)	.00	.03	.07

Model 1: predictor = pre-treatment parental attributions, outcome = post-treatment child conduct problems, mediators = post-treatment negative parental feelings and harsh discipline; Model 2: predictor = pre-treatment parental attributions, outcome = 3-month follow-up child conduct problems, mediators = post-treatment negative parental feelings and 3-month follow-up harsh discipline; Model 3: predictor = changes in parental attributions, outcome = 3-month follow-up child conduct problems, mediators = post-treatment negative parental feelings and 3-month follow-up harsh discipline; changes in parental attributions = post- minus pre-treatment parental attributions; paths defined in Fig. 1

*p-value < .05