New Iterative AM Estimation Procedure for Fitting the Simple Linear Measurement Error Models

Abstract

This article proposed a modified AM estimation procedure. The procedure uses the grouping estimators iteratively after dividing the sample into clus- ters. Then, the grouping AM procedure used to fit the structural relationship with measurement error considering there is no equation error model. The performance of the iterative grouping estimator is compared with the tra- ditional two group estimators. Simulation study showed that in terms of mean square error the proposed estimator is robustify the traditional two group estimator. A real data analysis for studying the relationships between happiness rate and the corruption perceptions index in the Arabs states is discussed.