Are there any books that go into this topic in depth, typically I have seen 1 or 2 paragraphs devoted to the topic. Maybe a source with a conceptual focus.
For instance some questions: Does the Effect Modifier have to be dichotomous or can it be continous? I think it can be continuous because of the following scenario: Say you are testing that Coffee lowers body weight only in people with low kcal intake. Therefore you have to test for interaction because Var_kcal's beta alone would not tell you if calories affects coffee's effect on Bodyweight - it would only tell you of Calories' effect on bodyweight. In this case you would have both product terms being continuous, so it is possible. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
