Need to consider variations within demographic groups when evaluating educational interventions

Reviewer: Molisa D. Derk

As computer science (CS) educators strive to increase recruitment and retention efforts, especially among underrepresented groups, many studies have been conducted to measure the response of students, divided among demographic groups, to various such efforts. Kumar makes the valid point, backed up by data, that measuring effects among simply defined groups such as gender, ethnicity, and race may be misleading. In order to obtain meaningful results, groups must sometimes be subdivided. For example, Kumar's data shows that Caucasian and non-Caucasian females measure very differently on the self-confidence factor scale, a result also seen in other studies. Simply dividing the students along either gender or racial lines would miss this potentially important point. However, the author seems to imply that the need for subdividing demographic groups has been largely ignored by researchers. This is not the case. In fact, education research frequently studies subgroups. Kumar does make the point that this practice is needed and should continue. Online Computing Reviews Service