A researcher randomly assigns 120 students to a mindfulness training program (X: 1 = training, 0 = control). They hypothesize mindfulness improves exam performance (Y) by reducing perceived stress (M). They run a simple mediation in SPSS using PROCESS Model 4 (5,000 bootstrap samples). Selected output is below:
Path a (M ~ X):
- X → M: b = -5.20, SE = 1.10, t = -4.73, p < .001
Path b and c' (Y ~ X + M):
- M → Y (controlling X): b = -0.40, SE = 0.12, t = -3.33, p = .001
- X → Y (direct effect c', controlling M): b = 1.10, SE = 0.88, t = 1.25, p = .214
Total effect c (Y ~ X):
- X → Y: b = 0.98, SE = 0.90, t = 1.09, p = .279
Indirect effect (a*b):
- Effect = 2.08
- BootSE = 0.74
- 95% bootstrap CI: [0.75, 3.65]
Which interpretation is MOST correct?
There is no mediation because the total effect (c) is not significant; therefore, mindfulness does not affect exam scores through stress.
There is evidence of a significant indirect effect of mindfulness on exam scores through stress; the indirect effect is significant because the bootstrap CI does not include 0, even though the total effect (c) is not significant.
This output indicates moderation (an interaction) because the effect of stress on exam scores changes depending on mindfulness condition.
Because c' is not significant, we can conclude full mediation (all of the effect of mindfulness on exam scores is explained by stress).