Two-sample Non-randomized Response Techniques For Sensitive Questions

In this article, we will show how to use the non-randomized model to compare the propor-tions of. people with sensitive characteristic in two independent populations. In order to achieve our goal, we derive the Wald test, Score test, the Likelihood Ratio test (the LR test) as well as their sample size based on the desired power. We use simulation to compare the performance of the three tests and the effects ofπ1, p and n on typeⅠerror and power. The results show that the effect ofπ1, p and n is not obvious for typeⅠerror. But we can get larger power with samller p and larger n. A real data set is used to illustrate the proposed methods.