The randomized response (RR) technique was first introduced by Warner (Ref. 1) to improve respondent cooperation and to encourage honest responses for estimating the population proportion n of individuals belonging to a sensitive group A. His method uses two randomized questions both inquiring about the respondent's status in relation to the sensitive characteristic. Another type of technique for solving the sensitivity problem is the unrelated question RR technique suggested by Simmon as recorded by Greenberg, et al. (Ref. 2). Instead of both questions referring to the sensitive characteristic A, one of them is replaced by a statement about a nonsensitive, innocuous characteristic Y. It has been verified helpful for reducing response error when potentially sensitive topics, such as the prevalence of illegal activities, incidence of domestic violence, sexual practice or other issues that are unethical or socially unacceptable are considered in finite population sampling surveys. A number of improvements of RR sampling have been suggested in the literature. See Chaudhuri and Mukerjee (Ref. 3) for a review.
展开▼
