## 8.8 Multicollinearity

• Occurs when some of the X variables are highly intercorrelated.
• Affects estimates and their SE’s (p. 143)
• Look at tolerance, and its inverse, the Variance Inflation Factor (VIF)
• Need tolerance < 0.01, or VIF > 100.
car::vif(mv_model)
##     FAGE  FHEIGHT
## 1.003163 1.003163
tolerance = 1/car::vif(mv_model)
tolerance
##      FAGE   FHEIGHT
## 0.9968473 0.9968473
• Solution: use variable selection to delete some X variables.
• Alternatively, use dimension reduction techniques such as Principal Components