Medium
Moderation Analysis
Free
In a moderation analysis with continuous predictors, a researcher mean-centers X (job demands) and W (autonomy) before creating the interaction X×W. Which statement best reflects why centering is commonly used?
A
Centering makes the interaction term statistically significant if it was not significant before.
B
Centering reduces multicollinearity between the interaction term and its components and makes main effects easier to interpret (as effects at the average level of the moderator).
C
Centering is required to meet the normality assumption for regression residuals.
D
Centering changes the overall model fit (R²) and the predicted values, so it should be avoided.