In research, which term describes distortion caused by other variables to both exposure and outcome?

In research, which term describes distortion caused by other variables to both exposure and outcome?

• A. Confounding
• B. Bias
• C. Interaction
• D. Random error

Answer: (A)

Confounding describes distortion that arises when a third variable is related to both the exposure and the outcome. Because this variable influences both, the observed association between exposure and outcome can be misleading, reflecting the effect of the confounder rather than a true effect of the exposure. A useful way to think about it is that a confounder is a variable that is associated with who gets exposed and also with the outcome, and it lies outside the causal pathway from exposure to outcome. For example, if studying the relationship between a medication and blood pressure, age could be a confounder if older people are more likely to use the medication and also have higher blood pressure independently of the medication. If you don’t account for age, you might wrongly attribute changes in blood pressure to the medication. Controlling for confounding is essential and can be done through study design (randomization, restriction, matching) or analysis (stratification, multivariable models). This distinguishes confounding from other concepts: bias is a broader term for systematic error, interaction means the effect of exposure changes across levels of another variable, and random error refers to random variability that affects precision.

Confounding describes distortion that arises when a third variable is related to both the exposure and the outcome. Because this variable influences both, the observed association between exposure and outcome can be misleading, reflecting the effect of the confounder rather than a true effect of the exposure.

A useful way to think about it is that a confounder is a variable that is associated with who gets exposed and also with the outcome, and it lies outside the causal pathway from exposure to outcome. For example, if studying the relationship between a medication and blood pressure, age could be a confounder if older people are more likely to use the medication and also have higher blood pressure independently of the medication. If you don’t account for age, you might wrongly attribute changes in blood pressure to the medication.

Controlling for confounding is essential and can be done through study design (randomization, restriction, matching) or analysis (stratification, multivariable models). This distinguishes confounding from other concepts: bias is a broader term for systematic error, interaction means the effect of exposure changes across levels of another variable, and random error refers to random variability that affects precision.