Which statement correctly describes confounding in community health research?

• A. Confounding is random error that cannot be controlled.
• B. Confounding is a bias that arises only from measurement error in exposure.
• C. Confounding is a bias that always inflates effect estimates.
• D. Confounding occurs when a third variable is associated with both exposure and outcome, creating a spurious association; it biases results if not controlled (via design or analysis).

Answer: (D)

Confounding in community health research happens when a third variable is linked to both the exposure and the outcome in a way that isn’t part of the causal pathway, so the observed association between exposure and outcome is distorted. Because this third variable influences both sides, it can create a spurious link that isn’t due to the exposure itself. This bias remains if you don’t account for the confounder through study design or analysis. It’s not simply random error, and it isn’t limited to measurement error; even perfectly measured data can be confounded if a relevant third variable is present. Also, confounding doesn’t always push the effect estimate upward; it can overestimate, underestimate, or even obscure the true association depending on how the confounder relates to both exposure and outcome. Control methods include design approaches like randomization, restriction, and matching, as well as analytical techniques such as stratification, multivariable adjustment, or propensity scores. A helpful example is when age is related to both the exposure and the outcome; if you don’t adjust for age, you might misattribute effects to the exposure that are actually due to age.