This workshop is an overview of the concepts behind mediation and moderation. It is not intended to be an in-depth statistical discussion of this topic, but a starting point for understanding how and when you might test these concepts using statistical modeling.
I’ve attached a Powerpoint from the presentation here (see notes on slides for more info). Video available upon request.
A few reminders from questions that arose during the presentation:
- It matters whether your variables are categorical or continuous for the various statistical tests you do to confirm mediation/moderation. Refer to other resources to choose the appropriate tests.
- If you’re trying to decide whether you’re testing mediation or moderation, draw a picture.
- If you think the relationship between X & Y exists because X predicts some other variable (Z) which predicts Y, then you’re probably interested in mediation
- If you think the relationship between X & Y is different for different groups (race/gender/ed level/etc.), then you might be interested in moderation
- Theory is the basis for all decisions you make in terms of what you want to test.
Access Powerpoint here: Mediation & Moderation
Tags: medation, model, moderation, sem, statistics, theory