Control Variables

In correlational research, a control variable might be labeled a confounding variable or nuisance variable that is “held constant” by statistical means. Suppose we want to know the relations between length of study time and scores on a test [Page 401]of American history, but we are worried that interest in history might be an alternate explanation of the association. If we allowed students to choose and report their own study times for the test, and we also measured the participants’ interest in history, we could adjust the relations between study time and test score by statistically holding constant scores on interest in history. In such a study, interest in history would be described as a control variable.