Hi guys. I am working with a doctoral student as a statistics consultant. However, I find myself at a loss for the current situation.

We have a relatively basic path model. Recently it was decided that we needed to include a moderating variable. I'm fine with modeling the basic aspect of the moderator, which is creating an interaction variable (e.g., X1, X2, and X1*X2). The difficulty I'm having is in the concept of mean centering. Can someone enlighten me as to how to do the mean centering process?

Another question:
Must you know of an interaction before using a moderator or is it acceptable to assume a moderator in you model without testing for interaction?

Thanks in advance for your help.