Social Contexts for Detecting Deception: Factors that Moderate the Effectiveness of Cognitive Load Approach Interviews

Cognitive Load Approach interviews are designed to improve deception detection accuracy by making an interview more difficult for liars than truth-tellers. However, factors such as the social context of an interview may reduce the feasibility of this interview approach. Individuals who perceive they are stereotyped as criminals may experience stereotype threat when interviewed by someone of a different ethnicity, resulting in high cognitive demand for both liars and truth-tellers. The similarity in demand may reduce lie-truth discrimination accuracy. These predictions were tested across two experiments. In Experiment 1, Hispanics and Whites were randomly assigned as liars or truth-tellers. They completed a mock-crime and were interviewed by a White interviewer. In Experiment 1a, Hispanic interviewees experienced more stereotype threat, arousal and cognitive load than Whites regardless of actual veracity. In Experiment 1b observers viewed videos of Experiment 1a interviewees and rated the degree that interviewees displayed behavioral cues related to lying or truth-telling. In Experiment 1c, another set of observers watched videos from Experiment 1a and made deception judgments. In Experiment 1b and 1c, the magnitude of behavioral differences between liars and truth-tellers was more similar for Hispanics than Whites, reducing lie-truth discrimination accuracy to about chance levels for Hispanics (44% accuracy) but not Whites (61% accuracy). In Experiment 1d, observers' lie-truth accuracy depended on an interviewee's working memory capacity; observers were less accurate detecting liars among White interviewees with higher working memory capacity, whereas observer discrimination accuracy was similar for Hispanics with high and low working memory capacity. Identical procedures as Experiment 1a-d were used in Experiment 2, but with a Hispanic interviewer. Hispanic interviewees reported experiencing more threat, but their arousal and cognitive load ratings did not significantly differ from Whites. Observers' ratings and lie-truth discrimination accuracy of interviewees did not differ significantly as a function of ethnicity or veracity. However, observers' lie-truth accuracy depended on the working memory capacity of interviewees; observers were more accurate discriminating between White interviewees with low than high working memory capacity whereas observer discrimination accuracy was similar for Hispanics with high and low working memory capacity. Implications of these results are discussed.