Table 3.

Regression coefficients, standard errors (SE), and model summary information for simple mediation models assessing a) the effect that the baseline TADS total score (X) exerts on the change in the BDI sum score (Y) through the BRS sum score (M) and b) the effect that the baseline TADS factor score (X) exerts on the change in the BDI sum score (Y) through the BRS factor score (M). The results are adjusted for two possible confounders: age at baseline (U₁) and gender (U₂)

a. Sum scores model
Antecedent	Consequent
	Mediator (BRS)				Dependent (BDI change)
		Coeff	SE	p		Coeff	SE	p
X (TADS)	a	−0.01	0.00	<.001	c’	−0.04	0.03	.22
M (BRS)		–	–	–	b	2.17	0.82	.01
U1 Age	a2	−0.06	0.03	.03	c´2	−0.83	0.38	.03
U2 Gender	a3	−0.50	0.10	<.001	c’3	1.19	1.51	.43
Constant	iM	4.60	0.46	<.001	iy	3.03	6.30	.63
		R2 =.18				R2 =.06
		F(21.21) = 293, p <.001				F(4.94) = 292, p <.001
		ab = −0.03; Boot SE = 0.01; 95% CI (boot) = −0.04 to −0.01
b. Factor scores model
Antecedent	Consequent
	Mediator (BRS)				Dependent (BDI change)
		Coeff	SE	p		Coeff	SE	p
X (TADS)	a	−0.24	0.06	<.001	c’	−0.80	0.67	.23
M (BRS)		–	–	–	b	1.80	0.64	.01
U1 Age	a2	−0.07	0.03	.03	c´2	−0.83	0.38	.03
U2 Gender	a3	−0.62	0.13	<.001	c’3	1.35	1.51	.37
Constant	iM	1.68	0.59	.01	iy	9.15	6.59	.17
		R2 =.14				R2 =.06
		F(16.17) = 295, p <.001				F(5.02) = 294, p <.001
		ab = −0.43; Boot SE = 0.16; 95% CI (boot) = −0.77 to −0.14