Expertise in Deception Detection Involves Actively Prompting Diagnostic Information Rather Than Passive Behavioral Observation

Abstract

In a proof-of-concept study, an expert obtained 100% deception-detection accuracy over 33 interviews. Tapes of the interactions were shown to N=136 students who obtained 79.1% accuracy (Mdn=83.3%, mode=100%). The findings were replicated in a second experiment with 5 different experts who collectively conducted 89 interviews. The new experts were 97.8% accurate in cheating detection and 95.5% accurate at detecting who cheated. A sample of N=34 students watched a random sample of 36 expert interviews and obtained 93.6% accuracy. The data suggest that experts can accurately distinguish truths from lies when they are allowed to actively question a potential liar, and nonexperts can obtain high accuracy when viewing expertly questioned senders.