Multicollinearity
A situation in multiple regression where two or more predictor variables are highly correlated with each other.
What is Multicollinearity?
Multicollinearity occurs when independent variables in a regression model contain overlapping information. While it doesn't reduce the predictive power of the model as a whole, it makes it mathematically difficult for the model to estimate the individual effect of each collinear predictor.
Why Multicollinearity Matters
It inflates the standard errors of the regression coefficients, making them highly sensitive to small changes in the model and often rendering previously significant variables statistically insignificant.
How to Interpret
Typically detected by looking at the Variance Inflation Factor (VIF). A VIF greater than 5 or 10 indicates problematic multicollinearity.
Example
Predicting a person's weight using both their 'height in inches' and 'height in centimeters' as independent variables would result in perfect multicollinearity, breaking the model.
Common Mistakes
- Discarding highly correlated predictors indiscriminately, potentially causing omitted variable bias if the discarded variable was theoretically critical.