NONRANDOMIZED RESPONSE MODEL FOR COMPLEX SURVEY DESIGNS
NONRANDOMIZED RESPONSE MODEL FOR COMPLEX SURVEY DESIGNS
Author(s): Raghunath Arnab, D. K. Shangodoyin, Antonio ArcosSubject(s): Economy
Published by: Główny Urząd Statystyczny
Keywords: complex survey designs; parallel model; randomized response; probability proportional to size; varying probability sampling
Summary/Abstract: Warner’s randomized response (RR) model is used to collect sensitive information for a broad range of surveys, but it possesses several limitations such as lack of reproducibility, higher costs and it is not feasible for mail questionnaires. To overcome such difficulties, nonrandomized response (NRR) surveys have been proposed. The proposed NRR surveys are limited to simple random sampling with replacement (SRSWR) design. In this paper, NRR procedures are extended to complex survey designs in a unified setup, which is applicable to any sampling design and wider classes of estimators. Existing results for NRR can be derived from the proposed method as special cases.
Journal: Statistics in Transition. New Series
- Issue Year: 20/2019
- Issue No: 1
- Page Range: 67-86
- Page Count: 20
- Language: English