ESTIMATORS OF ODDS RATIO REGRESSION PARAMETERS IN MATCHED CASE-CONTROL STUDIES WITH COVARIATE MEASUREMENT ERROR

This article studies estimators of the odds ratio and odds ratio regre ssion parameters in finely matched case-control studies containing a b inary exposure of primary interest and subject-specific covariates tha t are subject to measurement error. A retrospective logistic regressio n model for the binary exposure variable is used. The effect of measur ement errors on the conditional maximum likelihood estimator is determ ined. Three alternatives are considered: bias-corrected, functional, a nd ''transformation'' estimators. The asymptotic and small-sample prop erties of the three competitors are studied. The results are illustrat ed using data from a case-control study of diet and colon cancer.