A NOTE ON PROTECTION OF PRIVACY IN RANDOMIZED RESPONSE DEVICES

다국어 초록 (Multilingual Abstract)

We consider 'efficiency versus privacy-protection' problem concerned with several well-known randomized response (RR) devices to estimate pro­portion of people bearing a stigmatizing characteristic in a community. The literature of RR on respondent's privacy protection discusses only about response specific jeopardy measures. We propose a measure of jeopardy that is independent of the RR offered by the interviewee and recommend it for using as a technical characteristic of the RR device. For ensuring better cooperation from the interviewees this new measure that depends only on the design parameters of the RR devices may be disclosed to the respondents before producing the RR by implementing the randomization device.

참고문헌 (Reference)

1 "？Using randomized response from a complex survey to estimate asensitive proportion in a dichotomous finite population？" 94, : 37？-42, 2001a

2 "？Unbiased variance estimation for multi-stage designs？" 1975

3 "？On the degree of protection in randomized interviews？" 1976197？203

4 "？On randomized response survey for estimating a proportion？ Communications in Statistics" 1994

5 "？Estimation of a Proportion through Randomized Response？" 1976

6 "？Estimating sensitive proportions from unequal probability sampleusing randomized responses？" 17( (17(): 259？-270, 2001b

7 "？Efficiency versus protection in a general RR-model？ ScandinavianJournal of Statistics" 1977

8 "a survey technique for eliminating evasiveanswer bias？ Journal of the American Statistical Association" 1965

9 "Journal of the RoyalStatistical Society" 1994

10 "Journal of the AmericanStatistical Association" 1969

11 "Journal of the American Statistical Association" 1976

12 "Journal of Indian Society of Agricultural Statistics" 1992