Question

Which of the following is true about multicollinearity?

Answer 1

Answer:

d. It is best measured using the statistic variance inflation factor (VIF).

Explanation:

Multicollinearity is an important issue in multiple regression model, having many independent/ explanatory variables. Multicollinearity is the situation in which two or more independent variables are highly correlated. It is problematic because it increases the standard error of independent variable coefficient & undermines its statistical significance

Variance Inflation Factor [VIF] is a check & corrective measure of multicollinearity.  

• VIF as a multicollinearity check : It quantifies the correlation between one explanatory variable with other explanatory variables.VIF = 1 implies there is no multicollinearity (correlation between independent variables); VIF upto 5  implies there is moderate multicollinearity (correlation between independent variables). VIF > 5 implies high multicollinearity (correlation between independent variables)
• VIF as a multicollinearity correction : Calculating  $Var (\beta j )$ = σ^2 / $[TSS j (1 - R^2j)]$ ; where TSS = total sum of square of variable j , σ^2 =  j variance, R^2 j = R^2 from regressing all other independent variable on variable j