Applying Implicit Propositional Measures to Detect Faking in Personality-Related Scales: Reliability, Discriminating Power, and Classification Accuracy.

Abstract

Two implicit propositional measures designed to detect faking in personality-related scales were tested across four experimental studies. Study 1 (n = 116) included the Deception Relational Responding Task and Narcissistic Admiration and Rivalry Questionnaire as the faking-detector and target scale, respectively. Respondents were randomly assigned to faking or no-faking conditions. Faking respondents were instructed to appear opposite to their narcissistic profile, while no-faking had to respond honestly. In Study 2 (n = 133), the faking-detector was the Deception Implicit Association Test (Dec-aIAT), while the target scale and faking/no-faking instructions remained the same. In Studies 3 (n = 74) and 4 (n = 111), the faking-detector was again the Dec-aIAT, while the target scale was the Big Five Questionnaire-2. Faking respondents had to adhere to a desirable target profile (Study 3; faking-good) or its opposite (Study 4; faking-bad) while no-faking should respond honestly. Overall, the implicit measures showed adequate-to-excellent reliability, discriminating power, and classification accuracy.