Table 6.

Sequential mediation analyses examining association between group assignment, change in implementation variables, change in psychosocial variables, and change in outcomes.

	Group assignment → change in implementation variable a path		Change in implementation variable → change in psychosocial variable d path		Change in psychosocial variable → change in outcome e path
	Estimate (SE)	p-Value	Estimate (SE)	p-Value	Estimate (SE)	p-Value
Physical activity (n = 632)
Messages and self efficacy	.10 (.05)	.04	.06 (.02)	.01	.45 (.22)	.04
Messages and social support	.10 (.05)	.04	.28 (.03)	<.0001	.33 (.18)	.07
Opportunities and self efficacy	.30 (.04)	<.0001	−.08 (.03)	.01	.39 (.21)	.07
Opportunities and social support	.30 (.04)	<.0001	.33 (.04)	<.0001	.32 (.17)	.06
Pastor support and self efficacy	.17 (.03)	<.0001	.03 (.03)	.38	.47 (.22)	.03
Pastor support and social support	.17 (.03)	<.0001	.33 (.04)	<.0001	.32 (.17)	.06
Policy and self efficacy (n = 314)	.28 (.04)	<.0001	−.21 (.05)	<.0001	.75 (.39)	.06
Policy and social support (n = 314)	.28 (.04)	<.0001	−.13 (.05)	.01	.62 (.28)	.03
Healthy eating (n = 627)
Messages and self efficacy	.13 (.05)	.01	.18 (.02)	<.0001	.03 (.15)	.85
Messages and social support	.13 (.05)	.01	.32 (.03)	<.0001	.11 (.11)	.33
Opportunities and self efficacy	.04 (.05)	.37	.26 (.02)	<.0001	.07 (.15)	.62
Opportunities and social support	.04 (.05)	.37	.22 (.03)	<.0001	.13 (.09)	.16
Pastor support and self efficacy	.13 (.04)	.003	.14 (.03)	<.0001	.04 (.14)	.77
Pastor support and social support	.13 (.04)	.003	.33 (.03)	<.0001	.12 (.10)	.21
Policy and self efficacy (n = 316)	.53 (.04)	<.0001	.03 (.05)	.59	.38 (.28)	.17
Policy and social support (n = 316)	.53 (.04)	<.0001	−.03 (.06)	.60	.15 (.15)	.33

Note: If all three models (i.e. a, d, and e paths) were significant, there was significant mediation; p < .05 considered significant.