Hi all-
Can anyone tell me if you are modeling a dependent variable (Y) and have a continuous variable (x1) as an independent, but find that there is a mass at a certain value of X1 or that the relationship between Y and X1 changes at a certain value of X1, so you introduce a dummy variable (either straight 0/1 or maybe X1-c where c is the "change point") do you need to worry about multi collinearity?
Also, do you then interpret the coefficient on X1 like you do with interaction (it depends on the value of X1 you are looking at)?
Thanks very much!