Examining the efficacy of an iterative measurement error adjustment technique on the adequacy of a 2-level model

Abstract

This paper proposes an iterative technique to realistically adjust for the incidence of measurement errors in multilevel models using a 2-level model framework. The technique yields the expected measurement error adjustment benefits but shows that such benefits do not necessarily accrue when all perceived error-prone predictor variables in a model are simultaneously adjusted for errors.

Keywords: Multilevel Model, Measurement Errors, Coefficient of Variation, Predictor Variable, Model
Deviance.