Moderator
A variable that changes the strength or direction of the relationship between an independent and dependent variable.
What is Moderator?
A moderator defines the conditions under which an effect occurs. It specifies when or for whom the relationship between a predictor and an outcome is stronger, weaker, or absent. In statistical terms, this is typically tested as an interaction effect.
Why Moderator Matters
Identifying moderators is crucial for understanding the boundary conditions of theories and for recognizing that interventions may not work identically across all subpopulations or contexts.
Example
A therapy program (IV) reduces anxiety (DV) overall, but the effect is much stronger for younger adults than older adults. Age acts as a moderator of the therapy's effectiveness.
Common Mistakes
- Confusing moderation (when an effect changes) with mediation (how an effect occurs).
- Failing to center continuous variables before creating interaction terms, leading to severe multicollinearity issues.