Models for survey nonresponse and bias adjustment techniques

Abstract

Survey statisticians have been dealing with the issues of nonresponse in sample surveys for many years. Due to the complex nature of the mechanism, so far it has not been easy to find a general solution to this problem. In this paper, several aspects of this topic will be elaborated on: the survey unit nonresponse bias has been examined alternatively by taking response amounts which are fixed initially and also by taking the response amounts as random variables. An overview of the components of the bias due to nonresponse will be performed. Nonresponse bias components are illustrated for each alternative approach and the amount of bias was computed for each case.