Evaluation of the randomized response technique in a drinking survey.

Publisher Website
Google Scholar
Abstract
In the 1st session, 59 high-school students (grade 12, 36 girls) completed an anonymous questionnaire which included an item on whether they had consumed 50 or more drinks during the previous year. In the 2nd session, 10 days later, the randomized response technique was used: the same students were asked to answer 1 of 2 questions according to the result of a coin toss; 1 question was innocuous (does your telephone number end in an odd digit?); the other was the question of drinking 50 or more drinks. The experimenter did not know which question had been answered. In the 1st session, 63% of the students answered the question on drinking affirmatively, and in the 2nd session, 85% (estimated from the probability of the coin toss and affirmative answers). The traditional self-report method may not elicit the optimum number of affirmative responses to a sensitive question.

This publication has 0 references indexed in Scilit: