Reducing Socially Desirable Responses in Epidemiologic Surveys An Extension of the Randomized-response Technique

Background: Even though the validity of self-reports of sensitive behaviors is threatened by social desirability bias, interviews and questionnaires are widely used in epidemiologic surveys on these topics. Methods: In the randomized-response technique, a randomization device is used to determine whether participants are asked to respond truthfully or whether they are prompted to provide a prespecified response. In this study, the randomized-response technique was extended by using a cheating-detection modification to obtain more valid data. The survey was on the dental hygiene habits of Chinese college students. Results: Whereas only 35% of men and 10% of women admitted to insufficient dental hygiene when questioned directly, 51% of men and 20% of women attested to this socially undesirable behavior in a randomized-response survey. Conclusions: Given the considerable discrepancy between the results obtained by direct questioning and by using the randomizedresponse technique, we propose that this technique be considered for use in epidemiologic studies of sensitive behaviors.