A Study of Various Methods of Random Response Against a Sensitive Character when Dishonest Response can Occur

Estimating the proportion of population carrying some sensitive character has always been a challenge to the survey statisticians because exposure of such character ( tax evasion, carrying venereal disease or addiction to drugs etc.) may lead to social disapproval and other serious problems of an individual. Warner's Method (S.L.Warner, 1965) is the pioneering one to asses the proportion of population carrying sensitive character usually denoted by A . He deployed the random response technique. The method is innovative but highly imprecise for firstly, it assumes that all responses are always true which is not very pragmatic and secondly, due to loss of information since all 'yes' responses are not given against the possession of sensitive character but also against the innocuous question as well. The next improvement came following by the work of Simmons (W.R Simmons, 1967), B.G. Greenberg et al. with the introduction of unrelated question. In the present paper attempt has been taken to include the average dishonest response factor and to obtain the modified expressions for the proportion of sensitive character.