Department of Statistics, The Ohio State University
Statistics and Biostatistics Colloquium Series

We propose a class of semiparametric estimators in a general setting of functional measurement error models. The estimators are derived using estimating equations that are based on the semiparametric efficient score derived under the (possibly) incorrect distribution for the "unobserved measured with error" covariates. It is shown that such estimators are consistent and asymptotically normal even with misspecification and are efficient if computed under the truth. Under certain conditions, such estimators yield a closed form expression which coincides with the efficient score estimator known in literature. The methods are demonstrated with a simulation study of linear and quadratic logistic regression model with measurement error. Such estimators can also be used in inference for auxiliary parameters in semiparametric mixture models.