Hi,
Here is my regression model:
regress y x1 x2 x3_x4; /* x3_x4 = x3 - x4 */
Do I need to add x3 (omitted x4 for the perfect collinearity ) in my regression model? And why? Is it similar to the regression with an interaction which needs to contain the main effect?
Thank you!
JC
0 Response to Do I need to add main effect in a regression model with a Difference value between two variables?
Post a Comment