An important goal of many epidemiological studies is to estimate the magnitude of association between an exposure and an outcome. Exposure measurement error causes bias in such estimates of association and can be substantial. In this article, we describe the problem of exposure measurement error and its effects. We show how a simple hand calculation, in conjunction with validation study data and a calibration equation, can be used to correct estimates for the bias caused by exposure measurement error. Correcting estimates of association for measurement error helps researchers appropriately assess effect size.
Keywords: bias; calibration; epidemiologic measurement; regression analysis; validation study.
© 2014 Asian Pacific Society of Respirology.