Estimation of Population Proportion of Stigmatized Attribute Using Bayesian Approach

Abstract

Bayesian Approach to Randomized Response Technique has been a technique for estimating the population proportion, especially of respondents possessing stigmatized attributes such as induced abortion, use of drugs and tax evasion. In this paper, we propose Bayesian estimators of population proportion of a stigmatized attribute assuming Kumaraswamy and the generalised beta prior using life data on induced abortion. The newly proposed Bayesian estimators were validated numerically for a large interval of the designed values of the population proportion at different sample sizes. It was observed that the newly developed Bayesian estimators were more sensitive in capturing stigmatized attribute than the Bayesian estimator developed by Hussain & Shabbir (2012) for relatively small, moderate as well as large sample sizes.

Keywords

• Bayesian estimators
• alternative priors
• stigmatized attribute
• mean square error
• absolute bias