Too good to be true: Comparing lying habits as reported by credible vs non-credible responders

Abstract

Researchers typically rely on self-report when inquiring about individuals’ behavior. Yet, when measuring socially unacceptable behavior such as dishonesty, self-report is prone to bias, especially underreporting (i.e., “positive response bias” or “faking good”). This study aimed to examine the impact of supernormality, an extreme form of positive response bias, on self-reported lying behavior. A total of 777 participants completed an online survey, which included assessments of lying behavior and supernormality. Based on their supernormality scores, participants were categorized into a "supernormal group" (n = 223, 28.7%) and a "credible group" (n = 554, 71.3%). Our results revealed that the supernormal group reported a significantly lower lying frequency, higher unwillingness to lie due to negative internal factors such as experiencing guilt (i.e., negativity), being less influenced by external factors in their decisions to lie (i.e., lower contextuality), lower relational lying, and lower antisocial lying, compared to the credible group. These group differences in lying behavior are interpreted as a manifestation of the supernormal group’s heighted motivation to “fake good” and as underreporting of their lying behavior. Furthermore, when controlling for supernormality levels, subtle gender differences in lying behavior were revealed, with women reporting lower proficiency in lying, being less influenced by contextual factors in their decisions to lie and engaging in less antisocial lying than men. Overall, our results highlight the unreliability of self-report when investigating sensitive topics, such as lying, and the importance of screening for response bias (i.e., underreporting).