Statistics Colloquium

Location

Mathematics/Psychology : 106

Date & Time

May 8, 2015, 11:00 am12:00 pm

Description

Speaker
Prof. Bikas Sinha
Retired Prof. of Statistics
ISI, Kolkata, India

Title
Eliciting Information on Sensitive Features : Block Total Response Technique and Related Inference


Abstract
Randomized Response Technique [RRT] was first introduced and popularized by Warner in 1965. Since then survey sampling theoreticians and practitioners have contributed significantly in this area of survey methodological research. The idea is to be able to elicit 'truthful' response on sensitive feature(s) from the sampled respondents (out of a finite labeled population of respondents) so that eventually the population mean of the sensitive feature can be 'unbiasedly' estimated. Towards this, a novel technique was introduced by Raghavarao and Federar (1979) and it was termed 'Block Total Response' [BTR] Technique. We undertake various meaningful versions/generalizations of the BTR Technique, after a brief review of the literature in this direction. In the process, we also introduce empirical Bayes [EB] estimators.

Key Words: Sensitive Feature, Regular Feature, Qualitative and Quantitative Features, Randomized Response Technique, Block Total Response Technique, Balanced Incomplete Block Design, Binary Proper Equireplicate Block Design, Priors, EB Estimators