Nonresponse in stratified sampling: a mathematical programming approach

Abstract

In sampling theory the term nonresponse is used for not being able to obtain from some units selected in the sample. Among other reasons nonresponse may be due to the refusal to answer or due to evasive answers in response to a sensitive question. Warner (1965) presented the Randomized Response (RR) technique to estimate the proportion of respondents to a sensitive question without the knowledge of the respondents' personal status. This paper addresses the problem of optimum allocation in stratified sampling under RR model as an All Integer Nonlinear Programming Problem (AINLPP) in the presence of nonresponse. The solution to the formulated problem is obtained using optimization software.