Please help!


I have a question: in a hierarchical regression when subsequent variables (X3 and X4) that are added later on results in the original variable (X1) being non-significant. What if X3 turns out to be not-significant. So do we say that only X4 can significantly predict the depended variable but not X3 and that X3 is not correlated with X1 and X2?

Thank you.


New Member
The best way to do this is to perform a "Principal Component Analysis" through which one finds the linear combinations of the independent variables, in order, which analyse the variation of dependent var the best.
You cannot say X3 is uncorrelated to X1 or any variable which turned out to be insignificant as X3 is so.