ABSTRACT
Suppose a survey is required to deal with sensitive issues like the extent to which habits of drunken driving, tax evasion, gambling, etc., are prevalent in a certain community in a given time period. The entire survey need not be exclusively concerned with such stigmatizing items of query, but some of the structured questions in an elaborate survey questionnaire may cover a few specimens like these. It is likely that an investigator will hesitate to raise such delicate questions, and people when so addressed may refuse to reply or supply evasive or false answers. As a possible way out one may try to replace a direct response (DR) query by a randomized response (RR) survey. We discuss briefly how it can be planned and implemented and indicate some possible consequences.
