A Confounding Variable is an extraneous variable whose presence affects the variables being studied so that the results you get do not reflect the actual relationship between the variables under investigation. When conducting an experiment, the basic question that any experimenter is asking is: “How does A affect B?” where A is the probable cause, and B is the effect. Any manipulation of A is expected to result in a change in the effect.
For example, you want to study whether bottle-feeding (Cause) is related to an increased risk of diarrhea in infants (Effect). It would seem logical that bottle-fed infants are more prone to diarrhea since water and the bottle could get contaminated, milk could go bad, etc. But if you were to conduct this study, you would learn that bottle-fed infants are less likely to develop diarrhea than breast-fed infants. It would seem that bottle-feeding actually protected against the illness.
But in truth, you would have missed a very important confounding variable – mother’s education. If you take mother’s education into account, you would learn that better-educated mothers are more likely to bottle-feed their infants, who are also less likely to develop diarrhea due to better hygienic practices of the mothers. In other words, mother’s education is related to both the Cause and the Effect. Not only did the Confounding Variable suppress the effect of bottle-feeding, it even appeared to reverse it – confounding results, indeed!
This example illustrates the importance of identifying and controlling for possible Confounding Variables in any research study. A thorough review of available literature should help you do this.