This paper applies Kim and Warde's (2004) stratified Warner's randomized response model to Mangat and Singh's (1990) two-stage randomized response model. The proposed stratified randomized response model has an optimal allocation and a large gain in precision. Hence, the estimator based on the proposed method is more efficient than Kim and Warde's (2004) and Mangat and Singh's (1990) estimators under the conditions presented in both the case of completely truthful reporting and that of not completely truthful reporting by the respondents.
- Dichotomous population
- Estimation of proportion
- Randomized response technique
- Sensitive characteristics
- Stratified random sampling