Evaluating different approaches for multiple imputation under linear constraints

Abstract

"In this paper we will focus only on imputation methods for variables that are subject to one linear constraint. In a simulation study based on data from the German IAB Establishment Survey, we will evaluate five different multiple imputation approaches that address the linear constraints in different ways. The remainder of the paper is organized as follows: Section II describes the data used for the simulation study. Section III introduces the five different imputation approaches. In Section IV the simulation design is illustrated. Section V presents the simulation results. The paper concludes with some final remarks and ideas for future research." (Author's abstract, IAB-Doku) ((en))

Cite article

Drechsler, J. & Raghunathan, T. (2008): Evaluating different approaches for multiple imputation under linear constraints. (United Nations, Economic Commission for Europe. Working paper 25), New York, 12 p.