Approximations for Regression with Covariate Measurement Error

From MaRDI portal





The effect of covariate measurement error on maximum likelihood and maximum quasilikelihood estimates (MLE's or MQLE's) of regression parameters is considered. The naive estimates, obtained by ignoring error, are modified to account for it. The modified estimates can be computed explicitly from the naive ones, and have less bias for small error. They are particularly useful for Berkson error models. Examples illustrate their application to linear and nonlinear regression, errors in continuous covariates, and misclassification of discrete covariates. The estimates are applied to data relating air pollution to respiratory disease in children. Analytic methods and simulations are used to compare the estimates to others that have been proposed for dealing with covariate measurement errors. The basic idea is that the MLE or MQLE of the parameter \(\theta\) is a function \({\hat \theta}\)(\(\delta)\) of the size \(\delta\) of the error. When \(\delta =0\) and there is no error, \({\hat \theta}\)(0) is the usual MLE or MQLE. For \(\delta\) small, an improvement over \({\hat \theta}\)(0) is \(\theta^*={\hat \theta}(0)+\delta \partial {\hat \theta}(0)/\partial \delta\). In many cases of practical interest \(\theta^*\) can be computed in terms of the mean and variance of the error conditional on the erroneously observed covariates.




Cited in
(24)








This page was built for publication: Approximations for Regression with Covariate Measurement Error

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3825964)