This question is probably basic, but I'm rather stuck on it. I got a regression model with one main independent variable and quite a bunch of controls. Basically, I want to check if the effect of my main variable still remains in the same direction for a subset of my sample (dummy control). Running a separate regression only for the subset is not an option due to sample size. Can I somehow make use of interactions? From what I understood, the interaction term would only show if the effect of the main IV on the DV is stronger or weaker when the control variable rises in its value.