📸 Multicollinearity diagnostics in SPSS Linear Regression
What Is Multicollinearity?
Multicollinearity occurs when predictor variables in a regression are highly correlated with each other. This inflates standard errors of regression coefficients, making them unstable and difficult to interpret. In severe cases, it can reverse the sign of coefficients.
Key diagnostic: VIF (Variance Inflation Factor). VIF>10 (some say >5) indicates problematic multicollinearity. Tolerance = 1/VIF; Tolerance < .10 signals the same problem.
Checking VIF in SPSS
Step 1: Analyze → Regression → Linear. Add your DV and predictors.
Step 2: Click Statistics → check Collinearity diagnostics → Continue → OK.
📸 VIF output — all below 10, no serious multicollinearity
Remedies If VIF Is High
- Remove one of the highly correlated predictors
- Create composite variable (sum or average)
- Use PCA to create orthogonal components
- Apply Ridge regression or LASSO
APA Reporting
Multicollinearity diagnostics indicated acceptable VIF values (VIF=1.22–2.39) and Tolerance values above .10, suggesting multicollinearity was not a concern in this model.