An alternative randomized response model using two decks of cards

Abstract

In an attempt to obtain trustworthy answers from the respondents regarding a sensitive question in order to estimate the population proportion possessing that sensitive attribute, Odumade and Singh (2009) have suggested a randomized response model based on the use of two decks of cards. In this paper, a modification to the structure of the two decks of cards used by Odumade and Singh (2009) is proposed. The condition when the proposed model is more efficient than the Odumade and Singh (2009) model has been obtained. The proposed model can be easily adjusted to be more efficient than the Warner (1965), Mangat and Singh (1990), Mangat (1994) and Odumade and Singh (2009) models. Another advantage of the proposed model is that it can increase the respondents’ cooperation as they will be required to answer the sensitive question only once and not twice as in the case of the Odumade and Singh (2009) model.