Table 2. Mediation of the association between maternal depressive symptoms^a and whether young person studied at higher education by the young person's educational achievement, depressive symptoms, and locus of control (n = 9852).

Mediator		OR	CI	p	% mediated
Educational achievement at 14	Total effectb	0.84	0.78,0.92	p < 0.001	60.4
	Controlled direct effectc	0.93	0.87,1.01	p = 0.08
	Natural indirect effectd	0.90	0.88,0.93	p < 0.001
Educational achievement at 16	Total effect	0.81	0.73,0.89	p < 0.001	82.7
	Controlled direct effect	0.96	0.89,1.05	p = 0.39
	Natural indirect effect	0.84	0.80,0.88	p < 0.001
Young person's depressive symptoms at 16	Total effect	0.88	0.82,0.94	p < 0.001	6.0
	Controlled direct effect	0.88	0.83-0.95	p < 0.001
	Natural indirect effect	0.99	0.98-1.00	p = 0.12
Young person's locus of control at 16	Total effect	0.87	0.82-0.94	p < 0.001	34.4
	Controlled direct effect	0.92	0.85-0.98	p = 0.02
	Natural indirect effect	0.95	0.94,0.97	p < 0.001

^aExposure: number of times the mother scored 13+ on the Edinburgh Postnatal Depression Scale during the child's lifetime. Odds ratios therefore represent associations per additional occasion.

^bTotal effect is the effect of the exposure on the outcome. Note that paramed calculates this from the direct and indirect effects, so the estimated total effect differs slightly between models.

^cControlled direct effect is the effect of the exposure on the outcome while controlling for the mediator.

^dNatural indirect effect and is the effect of the exposure on the outcome that works through the mediator.