confound

Confound, as a verb, means to cause confusion or to mistake one thing for another. In scientific research, a confounder is a variable that influences both the exposure and the outcome, potentially producing a spurious association.

Etymology: confound comes from Old English confundan and from Latin confundere, meaning “to pour together” or “to mix up.”

In epidemiology and statistics, confounding occurs when the observed link between an exposure and an outcome is distorted by a third variable. A confounder must be associated with both exposure and outcome and must not lie on the causal path from exposure to outcome. For example, smoking is a common confounder in studies of coffee and cardiovascular disease, because it is related to coffee consumption and to heart risk.

Control methods include study design and analysis. Design-based approaches include randomization, restriction (limiting study participation to specific subgroups), and matching. Analysis-based methods include stratification, multivariable regression, and propensity score techniques. Even with these methods, residual confounding may remain due to imperfect measurement or unmeasured variables.

In everyday language, confound means to perplex or confuse; in research, it refers to a potential bias that researchers seek to prevent or adjust for in order to draw valid conclusions.