Faking good and bad on self-reports versus informant-reports of Dark Triad personality

Abstract

Research consistently demonstrates that people can distort their responses on self-report personality tests. Informant-reports (where a knowledgeable informant rates a target's personality) can be used as an alternative to self-ratings. However, there has been little research on the extent to which informants can distort their responses on personality tests (or their motives for response distortion). The current study examines the effects of experimentally induced response distortion on self- and informant-reports of the Dark Triad. The participants (N = 834 undergraduates) completed Dark Triad measures in a 2 × 3 between-person design crossing format (self- vs. informant-report [imagined friend]) with instruction condition (answer honestly, look good, or look bad). “Look good” effects were significant for both self-reports (d = −1.22 to 1.42) and informant-reports (d = −1.35 to 0.62). “Look bad” effects were also significant for both self-reports (d = −0.56 to 3.58) and informant-reports (d = −0.55 to 3.70). The Five Factor Machiavellianism Inventory results were opposite to hypotheses, but Dirty Dozen Machiavellianism results were as expected. We conclude that people can distort Dark Triad scores for themselves (self-report) and on behalf of someone else (informant-report). We discuss the relevance of our findings for self- and informant-report assessment in applied contexts.