Hello Matthew,

The answer depends on whether the proposed moderator is thought to interact with one or both of the IVs.

If just one IV (say, X1), then you're in effect regressing the DV (Y) on:

Y = b0 + b1X1 + b2X2 + b3X1*W + b4X1*X2

If both (X1, X2), then you're regressing Y on:

Y = b0 + b1X1 + b2X2 + b3X1*W + b4X1*X2 + b5X2*W + b6X1*X2*W

Hayes, in his text, Introduction to Mediation, Moderation and Conditional Process Analysis: A Regression Approach (3rd ed.), doesn't give any examples of moderation involving more than one IV (X). There may be a way to coax the PROCESS macro to handle either or both models; I don't use it (see my preferred options below) enough to suggest how to do so.

Either of the above models may be estimated via ordinary least squares regression or a path model using an SEM program.

Good luck with your work.