Confounding variables are factors that can obscure or distort the true relationship between the independent and dependent variables in a study. In the context of cancer research, these variables can make it difficult to determine the actual cause-effect relationships between potential risk factors and cancer outcomes.