Manual
Moderation Analysis
The moderator variable is one that affects the relationship between two others: the strength
or direction of the relationship between the two vars is affected by the moderator
(moderation is the same as the interaction effect in ANOVA).
1. Draw a conceptual model
2. Look at the nature of the moderator:
1. If it’s categorical > interaction effect in ANOVA
2. If it’s continuous > use PROCESS
3. Check assumptions by running a multiple regression analysis
1. For the IVs, you include the IV, moderator and interaction term you create yourself
under compute (IV*Mod) as independent variables (predictors) in one block.
2. Include the DV as a DV
3. Check output for assumptions
1. Cook’s distance
2. Centered leverage value
3. Mahalanobis distance
4. Residuals
5. Multicollinearity
6. Normality of residuals
7. Durbin-Watson
8. Heteroscedasticity
9. Linearity
10. Bootstrap in PROCESS when not all assumptions are met.
4. Analyze > Regression > PROCESS
1. Outcome variable to Y
2. Predictor variable to X
Moderation Analysis
The moderator variable is one that affects the relationship between two others: the strength
or direction of the relationship between the two vars is affected by the moderator
(moderation is the same as the interaction effect in ANOVA).
1. Draw a conceptual model
2. Look at the nature of the moderator:
1. If it’s categorical > interaction effect in ANOVA
2. If it’s continuous > use PROCESS
3. Check assumptions by running a multiple regression analysis
1. For the IVs, you include the IV, moderator and interaction term you create yourself
under compute (IV*Mod) as independent variables (predictors) in one block.
2. Include the DV as a DV
3. Check output for assumptions
1. Cook’s distance
2. Centered leverage value
3. Mahalanobis distance
4. Residuals
5. Multicollinearity
6. Normality of residuals
7. Durbin-Watson
8. Heteroscedasticity
9. Linearity
10. Bootstrap in PROCESS when not all assumptions are met.
4. Analyze > Regression > PROCESS
1. Outcome variable to Y
2. Predictor variable to X