Abstract Intentional misreporting of sensitive behaviors may occur due to sensitivity bias. Interview-led surveys present an additional challenge: respondents may be hesitant to reveal sensitive behaviors directly to the interviewer. This article presents a novel method—the Colorbox—that attempts to resolve this particular issue. Instead of answering the interviewer directly, respondents reply through PIN codes that correspond to their desired answer. Their answers remain confidential to the interviewer, while researchers are able to decrypt the answers. Such a method performs well in contexts in which the main barrier to providing truthful responses is the lack of confidentiality with respect to the interviewer. This study presents two different case studies (the measurement of child labor among cocoa farmers in Côte d’Ivoire and unprotected sex among female sex workers in Senegal) to illustrate that this method can be implemented for both binary and categorical outcomes. The results show that the majority of respondents demonstrated an understanding of the methodology and perceived it to offer greater confidentiality than direct elicitation with respect to the interviewer.
Lépine et al. (Fri,) studied this question.