The impact of confidentiality assurances on participants’ responses to sensitive questions

Privacy, anonymity, and confidentiality are issues often addressed when collecting sensitive information (Tourangeau, 2017). Despite the importance of the concept, there is no consensus on what constitutes a sensitive question (Tourangeau, Rips, & Rasinski, 2000). According to Tourangeau and Yan (2007), a question is considered sensitive when it asks for a socially undesirable answer, requesting that the respondent admits that he or she has violated a social or legal norm. Due to the potential consequences of admitting some socially undesirable beliefs, behaviors, or activities (e.g.,social or legal sanctioning), sensitive questions are often associated with systematic measurement error (Groves et al., 2004). For instance, research has shown that sensitive questions have comparatively higher item nonresponse rates than nonsensitive questions, and they are regarded as one of the major causes for socially desirable answers in surveys(Diekmann, 2003; Krumpal, 2013; Na¨her & Krumpal, 2012; Tourangeau et al., 2000). Over the years, numerous methodological experiments have been conducted to analyze the impact of several factors on reducing measurement error in sensitive questions. Survey mode has been found to be one of the most important factors in explaining social desirability bias in sensitive questions (Holbrook, Green, & Krosnick, 2003; Tourangeau & Yan, 2007; Turner et al., 2005). It has been found that, in general, when interviewers collect the data or when they are present during data collection, responses are more socially desirable than in self-administered surveys (Kreuter, Presser, & Tourangeau, 2008; Tourangeau & Yan, 2007)