Modification of Unrelated Question Randomized Response Model*
Noodchanath Kongchouy and Surin Khanabsakdi*** Author for corresponding; e-mail address: sciskhnb@chiangmai.ac.th
Volume: Vol.29 No.1 (APRIL 2002)
Opinion
DOI:
Received: 15 October 2001, Revised: -, Accepted: 25 March 2002, Published: -
Citation: Kongchouy N. and Khanabsakdi S., Modification of Unrelated Question Randomized Response Model*, Chiang Mai Journal of Science, 2002; 29(1): 49-52.
Abstract
Sensitive questions could be considered too personal for the respondent to answer truthfully or, indeed answer at all. Sample surveys using direct questions always show biased estimators in the results. A randomized response technique was suggested by Warner in 1965. Five years later, Greenberg et al. developed the technique called "unrelated question randomized technique". The two techniques are always used with qualitative data. The technique used in this paper is modified from the unrelated question randomized technique. Two sensitive questions are asked, one in positive form and the other in negative form. Respondents draw one question by using a randomized device and then answer the question truthly either as "Yes" or "No". Statistics and probability are used to find two unbiased estimators of the true proportions.