When to use this
Use mediation for a hypothesised indirect pathway and moderation for a hypothesised interaction.
Why it matters
These models answer different questions and require theory established before interpreting conditional effects.
Step-by-step instructions
1. Define X, mediator M or moderator W, and outcome Y.
2. Confirm variable coding and temporal or theoretical ordering.
3. Choose the corresponding procedure.
4. Select bootstrap or interaction settings offered by the workspace.
5. Run the analysis.
6. Interpret paths, intervals, and conditional effects without causal overstatement.
Expected outputs
Path estimates, bootstrap intervals or interaction effects, diagnostics, and provenance.
How to interpret the results
An indirect-effect interval describes model-based evidence. A moderation term indicates that the estimated X–Y relationship varies with W.
Common mistakes
Do not call a cross-sectional indirect effect causal. Do not interpret only component-path p-values or omit the interaction hierarchy.
Was this article helpful?
Your response helps the methodology team improve this documentation.