Collinearity

Collinearity and Multicollinearity

Collinearity is "a condition that exists when two predictors correlate very strongly" (Meyers, Gamst, & Guarino, 2017, p. 189).

Multicollinearity is a condition that exists when "more than two predictors correlate very strongly" (p. 189).

If the predictor variables are too strongly correlated with one another the researcher will have a challenging time estimating the relationship between the dependent and predictor variables. SPSS is capable of providing a diagnostic to find out if multicollinearity exists within your data. This feature is able to determine the tolerance, or amount of variance in the individual variable not explained by the other predictor variables.

(Muijs, 2011, pg. 155-157)

contributed by Joseph W. Sullivan

Steps for How to Detect Multicollinearity in SPSS:

1. Click "Analyze"

2. Then Select "Regression"

3. Click "Linear"

4. Put all the IV's in the IV section and then move ONE IV into the DV box.

5. Uncheck all boxes in "Statistics" except for "Collinearity Diagnostics"

6. Click "Ok"

• The output should indicate if there is a VIF. If the VIF is above 3 there is likely multicollinearity issues, and if it is above 10 you are highly likely to have multicollinearity issues.

Detecting Multicollinearity in SPSS [1]

contributed by Britany Kuslis, WCSU Cohort 8

References:

Gaskin, James, director. Detecting Multicollinearity in SPSS. YouTube, YouTube.com, 26 Mar. 2011, www.youtube.com/watch?v=oPXjQCtyoG0.

Meyers, S., Gamst, G., & Guarino, A.J. (2017). Applied multivariate research: Design and interpretation. Thousand Oaks, CA: Sage Publications.