Statistics 2- Regression Diag. Multicollinearity

check for multicollinearity

What is the 5th step in the model building process?

multicollinearity

What exists in virtually all multiple regression models

multicollinearity

What is a condition that occurs when independent variables are correlated with one another?

One or both of the correlated independent variables must be removed from the model

How do you eliminate multicollinearity?

5.0

If any value of VIF exceeds ____ enough correlation exists between the independent variables to claim that multicollinearity is present in the regression model

variance inflator factor (VIF)

Multicollinearity can be detected using what?

1. sample coefficient may be far from the actual population parameter 2. T-Statistic will be small 3. interpretation of the coefficients

What are the 3 adverse affects of multicollinearity?

Above .9 not allowed Between .8 and .9 should be avoided Between .7 and .8 tolerable Below .7 preferred

What are the rules about correlations?

collinearity or intercorrelation

What are two other names for multicollinearity?

Start with the one with the highest VIF if still exists repeat

When eliminating multicollinearity how do you know which independent variable must be used and what do you do if it still exists?

only when two or more independent variables are highly correlated

When is the only time the problem of multicollinearity becomes serious?