AN IMPROVED QUALITATIVE RANDOMIZED RESPONSE MODEL UNDER RANKED SET SAMPLING.

This study introduces a new qualitative randomized response technique for gathering trust-worthy sensitive data using a perfect (imperfect) ranked set sampling design. This model takes into account the sizes (weights) of various balls on the randomizing device to choose one of the two questions (out of which one is sensitive); so the respondent feels, liberty in choosing balls. In this way, the likelihood of an honest response increases. It is established that the procedure is easy to apply and the estimate under the proposed model is more precise than the original Warner's randomized response model. In addition, an application to real medical data is considered. Finally, a cost analysis of the proposed model is also presented. [ABSTRACT FROM AUTHOR]