Class(es) of factor-type estimator(s) in presence of measurement error

Abstract

When data is collected via sample survey it is assumed whatever is reported by a respondent is correct. However, given the issues of prestige bias, personal respect and honor, respondents' self-reported data often produces over-or under-estimated values as opposed to true values regarding the variables under question. This causes measurement error to be present in sample values. This article considers the factortype estimator as an estimation tool and examines its performance under a measurement error model. Expressions of optimization are derived and theoretical results are supported by numerical examples. © 2012 JMASM, Inc.