Some Basic Results on the Extension of Quasi-Likelihood Based Measurement Error Correction to Multivariate and Flexible Structural Models.
Collaborative Research Center 386, Discussion Paper 196
Quasi-score equations derived from corrected mean and variance functions allow for consistent parameter estimation under measurement error. However, the practical use of some approaches relying on this general methodo\-logical principle was strongly limited by the assumptions underlying them: only one covariate was allowed to be measured with non-negligible error, and, additionally, this covariate had to be conditionally independent of the other covariates. This paper extends basic principles of this method to multivariate and flexible models in a way that, on the one hand, retains the neat statistical properties, but on the other hand, manages to do without the restrictive assumptions needed up to now.