In a cohort study that found an association between alcohol consumption and bladder cancer, 20,000 middle-aged men were asked about their drinking habits and then tracked for five years to watch for the development of cancer. Researchers asked the men about their diet and exercise habits, but not about smoking. What is the most likely source of error in this study?

A. Reporting bias

B. The placebo effect

C. Selection bias

D. Confounding variables