When to use it
- •You have a theoretical reason to believe X causes M, which in turn causes Y.
- •All three variables are measured on continuous (or near-continuous) scales.
- •You want to quantify how much of the total effect travels through the mediator.
Assumptions to check first
- •A defensible causal ordering — mediation is a claim about process, and cross-sectional data cannot establish temporal order on its own.
- •No unmeasured confounding of the M→Y relationship (the assumption most often violated and least often acknowledged).
- •Linear relationships between each pair of variables, and normally distributed residuals.
What to report
- •The indirect effect (a×b) with a bootstrap confidence interval — this is the test. Report the number of bootstrap resamples.
- •The direct effect (c′) and the total effect (c).
- •The a and b paths individually, with standard errors.
APA 7 example
The indirect effect of leadership on engagement through trust was significant, ab = 0.24, SE = 0.06, 95% bootstrap CI [0.13, 0.37], based on 5,000 resamples. The direct effect remained significant, c′ = 0.31, p < .001, indicating partial mediation.
Numbers are illustrative — the Advisor generates this sentence with your own results.
The mistake reviewers catch
A significant indirect effect is not proof of a causal mechanism. Baron and Kenny's original "steps" approach is also outdated — a bootstrapped confidence interval around the indirect effect is now the standard test, and the total effect does not need to be significant for mediation to exist.
Run it on your own data
Paste or upload your dataset — nothing leaves your browser. The Analysis Advisor checks the assumptions above, runs mediation, and drafts the APA Methods and Results text with your numbers in it.
Open the Analysis Advisor →Related analyses
Frequently asked questions
When should I use mediation?
Does X affect Y through M? You have a theoretical reason to believe X causes M, which in turn causes Y.
What do I need to report for mediation in APA 7?
The indirect effect (a×b) with a bootstrap confidence interval — this is the test. Report the number of bootstrap resamples. The direct effect (c′) and the total effect (c). The a and b paths individually, with standard errors.
What is the most common mistake with mediation?
A significant indirect effect is not proof of a causal mechanism. Baron and Kenny's original "steps" approach is also outdated — a bootstrapped confidence interval around the indirect effect is now the standard test, and the total effect does not need to be significant for mediation to exist.