Missing Data and Estimating Equations
DescriptionEstimating equations provide a general framework for finding estimators and studying their properties in various kinds of statistical models. Many of the common estimators are covered by the theory of estimating equations. The problem of missing data is common in a variety of observational studies but so far there is no literature linking estimation equations with missing data. The principal intellectual objective of this research is to develop a unified framework for handling missing data in estimating equations under the assumption that the missing pattern is random. The project will introduce a mechanism whereby one mitigates the effects of missing data through a reformulation of the estimating equations, imputed through a semi-parametric procedure. The proposed mechanism includes some existing imputation mechanisms as special cases. The project will also explore and inter-relate an array of relatively new estimation and inference solutions for estimating equations and investigate their properties. A second extension will consider the so-called non-ignorable missing data mechanism and study an estimation approach that makes full use of the available auxiliary information.
|Effective start/end date||1/09/07 → 24/03/09|