Suppose we are regressing Y = aX + bD + cXD. a and b are main effects of variables X and D, and c is the interaction effect. Assume both X and D are continuous variables.
How to interpret the effect of D? Clearly we have to use both main effect b and the interaction effect c, but how to do it? I understand that if D was an indicator variable (dummy), this could have been much easier.
However if D is a continuous variable, it seems more difficult as the units of b and c are different.