Evaluating compliance with a computer assisted randomized response technique: a qualitative study into the origins of lying and cheating

This paper presents the results of a phased qualitative study into the possibilities of a combination of computer assisted self-interviewing with the randomized response technique to decrease evasive answer bias when sensitive topics are studied. As a sensitive topic, compliance with rules concerning the right to receive welfare was studied. In total, 18 respondents (age 21– 63, 12 females, 6 males) filled out the questionnaire and were interviewed afterwards. In the first phase, 11 respondents did not follow the randomized response instructions, but after the questionnaire was adapted, cheating (i.e., not operating according to randomized response rules) diminished to nil in the second phase. Most respondents trusted the research setting enough to give incriminating answers to sensitive questions, but lying (i.e., not giving an accurate answer to a question) could not be totally prohibited. Recommendations are formulated for researchers who want to set up their own computer assisted randomized response questionnaires. 2004 Elsevier Ltd. All rights reserved.