In the context of a scientific study such as this one, a confounding factor is one that has influence on both the exposure and event variable(s), which may lead to over- or underestimation of the direct relationship between them (if any).
For instance: In this example, researchers may have had reason to believe that male gender is associated with both higher risk of obesity (the exposure variable) and adult-onset asthma (the event variable). If gender is not taken into account, one may claim that the finding of an association between obesity and asthma is simply an artifact due to the high proportion of male patients (likely to present with both). <em>Controlling</em> for that variable (such as by matching, as in this example) allows researchers to test for this hypothesis.
