AN ALTERNATIVE MODEL FOR RANDOMIZED RESPONSE TECHNIQUE

Abstract

Response to sensitive questions usually leads to either response bias or refusal bias. Recently, two models called: the simple model and the crossed model were proposed of which the crossed model was tested to be better than the simple model. But, it was later discovered that there is the need to redefine the model as its applicability was not strong enough. This motivated developing an alternative model for estimating the population proportion of respondents belonging to two sensitive attributes at a time which is the focus of this study. In addition, the variances and the relative efficiency of the developed model were done using R programming language to examine the performance of the developed model over the existing ones. The result shows that the developed model proved to be more efficient than the conventional model, of a note in its applicability. Therefore, a primary survey was carried out on students of Federal University of Technology, Akure, Ondo State of Nigeria to estimate the population proportion of those with the sensitive character Examination Malpractice, Premarital Sexual Intercourse and both. From the result, the estimate of the proportion of examination malpractice is 0.188 that of premarital sex is 0.252 and that of examination malpractice as well as premarital sex is 0.110. This implies that a victim of examination malpractice is likely not a victim of pre-marital sexual intercourse but a victim of sexual intercourse is likely to have been involved in examination malpractice before. The analysis indicated that the randomized response technique (RRT) is more efficient than the direct question (DQ) approach.