Misclassification

Abstract

Misclassification refers to the false characterization of a study characteristic due to measurement error. Information regarding the procedures used to measure the study data helps to infer whether misclassification is likely to have occurred and the suspected type of misclassification. Non-differential misclassification arises from non-systematic error in measuring the study data and in most instances leads to observing a relative risk that is closer to 1.0 than that obtained under ideal measurements (bias toward the null). Differential misclassification arises from systematic error in measuring the study data, the impact of which depends on the specific pattern of measurement error that has occurred.