Traditionally, educational interventions in Computer Science have been studied for their effect on entire classes, or specific demographic groups. But, in our studies, we have found that often, significant interactions exist among demographic groups. Treating demographic groups as homogeneous groups when evaluating educational interventions in Computer Science could miss subtle interactions among the groups.