Confounding occurs when the relationship between the exposure and outcome is distorted by a third variable, known as a confounder. This can lead to erroneous conclusions about causality. Epidemiologists use various methods, such as stratification and multivariable modeling, to adjust for confounders and reduce the risk of confounding errors.