Examining DIF in Perceived Anonymity When Sensitive Items are Endorsed

There has been a budding trend in survey research to define anonymity as a subjective perception instead of an objective circumstance. This study provides the first systematic exploration of anonymity perceptions in a Web-based survey context using an IRT approach. Keeping characteristics of the survey administration constant, we used DIF analyses to demonstrate that endorsements of sensitive items appeared to influence responses to a perceived anonymity measure. We also showed that sensitive items did not display DIF based on scale scores for perceived anonymity. Theoretical and methodological implications are discussed. Anonymity in surveys has been a topic of empirical inquiry in industrial and organizational psychology that has spanned decades of research (e.g., Dunnette & Heneman, 1956; Booth-Kewley, Larson, & Miyoshi, 2007; Lautenschlager & Flaherty, 1990). The underlying influence of anonymity on respondent behavior is predicated upon the idea that when respondents are anonymous, they will be more forthcoming in their responses to questions asked of them on a survey instrument (Corey, 1937; Podsakoff, Mackenzie, Lee, & Podsakoff, 2003). This assumption is considered axiomatic enough to b