I am proposing a study to find out whether cognitive restraint, disinhibition and hunger predict weight gain in pregnancy. However, I know from previous research that this effect differs depending on BMI category.

Am I correct in saying I could use a hierarchical regression analysis, where I enter BMI (as dummy coded data) in the first stage, and enter the other three IV's in the second stage.

My aim is to be able to see how the three eating IV's predicted weight gain in underweight, overweight and normal participants. Would this be the correct analysis to choose.

Many thanks in advance!