Asking Sensitive Questions Indirectly

Abstract

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Biometrika Trust is collaborating with JSTOR to digitize, preserve and extend access to Biometrika. SUMMARYWe propose a new method for conducting surveys on sensitive topics which does not require direct answers from the respondents. It overcomes some of the difficulties associated with traditional randomized response techniques and is applicable to both qualitative and quantitative characteristics. Data collected using the proposed method are distributed according to a mixture distribution and the problem reduces to that of estimating a mixing proportion or a mixing distribution. The method can be extended easily to allow multiple trials. This enables us to improve the efficiency of our estimator without sacrificing respondent protection and at no increase in sampling cost.

A note on the comparison of some randomized response procedures

“…This model generalizes in some sense the idea of Kuk (1990): if an 'R1'/'R2' response could be perceived more protective than a 'Yes'/'No' response by the interviewed subjects, giving an integer between 1 and K could be perceived even more protective. Like in the model of Kuk (1990), the PP-measures here considered are not able to quantify this additional perceived protection.…”

“…Moreover, a major criticism of the model proposed by Chang et al (2004) when seeking Institutional Review Board approvals is that it is somewhat unethical to ask a respondent to give untruthful response. (3) The family of models proposed by Kuk (1990) is better than the Simmons' family in terms of efficiency and privacy protection. From an empirical point of view, the study of van der Heijden et al (2000) shows that Kuk's procedure seems to perform slightly better than the forced-response procedure and markedly better than face-to-face direct questioning and computer assisted self-interviewing.…”

“…4.4. The Kuk (1990 In the procedure proposed by Kuk (1990) each respondent generates two binary outcomes according to two Bernoulli distributions with known parameters p K 1 and p K 2 , respectively. If the respondent has attribute A, he/she reports the outcome of the first random device.…”

“…Proposition 5. For p K 1 < p K 2 , ∈ (0, 1) and for each Simmons' model, there exists at least one Kuk (1990) model with equal L-level of protection and higher efficiency.…”

The Analysis of Randomized Response Sum Score Variables

“…Similar models have been presented for other RR designs such as the unrelated question design (Horvitz et al. , 1967), the forced response design (Boruch, 1971) and the Kuk design (Kuk, 1990).…”