In this video, we look at the difference between using a control variable, using a subsample, and using separate regression models on subsamples. Both conceptually and empirically, these are different approaches. We look at associations, generalization, and comparisons between coefficients when populations may vary on unmeasured aspects.

00:00 Intro