Confounding is a variable that may impact on the relationship between two other variables, making it difficult to establish a clear causal relationship. Researchers use statistical control techniques to reduce or eliminate the effect of confounding variables in research designs.