how to explian the insig. main effect without sig. coefficient of interaction?

#1
I am doing the research using the panel data. My confusions follow.
Y=x1+x2: X1 sig, but X2 insig.
however, when interaction is put into the model
Y=X1+X2+X1X2, all (X1, X2, X1*X2) have significant coefficients.
Anyone can help to explain the insig. X2 in first model?
really appreciate!!
 

kiton

New Member
#4
Why would you expect x2 be significant in model 1? Is there a theory or something about that? Now, in case of model 2, in the presence of an interaction the effect of x2 is the so-called conditional effect at the mean value of x1. As such, your results provide evidence that x2 could be significant in case x1 = mean -- i.e., x2 depends on x1 (or x1 moderates the relationship b/w x2 and y).
 
Last edited: