11.1.1 Dependent and Independent Variables The dependent variable is the outcome that researchers are trying to explain, In the above examples, that would be colon cancer, the cognitive ability of young children, and lifetime wages. An independent variable, or explanato- ry variable, is a factor that researchers believe has some causal relationship with the outcome of interest. Regression analysis alone cannot prove causali- ty, as will be explained in greater depth later in this chapter, However, it can prove a statistical correlation that can be used to help confirm causality when combined with other kinds of theoretical or empirical evidence. A control variable is an independent variable that is likely to have an effect on the outcome of interest but is not the primary focus of inquiry; researchers must include it in the regression analysis to improve the validity of the results. Suppose that researchers were trying to determine whether playing golf has a positive impact on health. Their specific outcome of inter- est is heart disease. Are people who play golf regularly more or less likely than nongolfers to suffer from heart disease? A simple comparison of the cardiac health of golfers compared to nongolfers would almost certainly find that golfers are less healthy. Does that mean that golf is dangerous? Should doc- tors be warning their patients to stay off the course? Of course not. You may have already figured out the flaw in this analysis: golfers tend to be much older on average than nongolfers since golf is a popular retirement sport. So a sim- ple comparison of golfers to nongolfers is really just comparing the cardiac health of older people to younger people; we wouldn't be surprised to find that heart disease is positively associated with age. In this case, researchers would have to include age as a control variable. The subsequent analysis would quantify the relationship between heart disease and playing golf, con