A MODIFIED EFFICIENT RANDOMIZED RESPONSE MODEL

Abstract

In this paper, the problem of estimating the population total of a sensitive quantitative variable has been considered. Odumade and Singh (2009) have given a RR model to estimate the population total of sensitive quantitative variable. We have given a modified RR model. It is noticed that the RR model envisaged by Odumade and Singh (2009) suffers with a difficulty that the optimum value of the constant involved in their RR model depends on the value of unknown parameter of the study variable, which lacks the utility of their RR model. To overcome such a difficulty Odumade and Singh (2009) have also suggested another estimator of the population total which depends on two RR models based on two independent random samples 1s and 2s
from the population  using the sampling design  1sp and  2sp respectively; and also on a relation between the two constants involved in the two RR models. We should add here that drawing the two independent samples from a population constructing two RR models resultantly obtaining an estimator for a population total Y increase the cost of the survey. Keeping this in view we have suggested a RR model and developed an estimator for population total based on a single sample. The proposed RR model is free from the difficulty involved in Odumade and Singh (2009) model. We have shown theoretically that the proposed randomized response model is more efficient than Bar-Lev et al (2004) and Odumade and Singh (2009) RR
models. This fact has also been supported through numerical illustration