If VIF value from a Regression model is 10.2, what should the Black Belt do next?

OPTIONS

• Separate the variables and conduct a Simple Linear Regression
• Drop the predictor that is causing the high VIF and then re-do Regression
• Live with the variable having high VIF and hope that it is a model adequacy issue
• Try doing a Curvilinear Regression

ANSWER

Drop the predictor that is causing the high VIF and then re-do Regression

EXPLANATION

The VIF model being so high, the Black Belt should drop the predictor that is causing the high VIF and then re-do Regression.