J. Pete Blair

The gap between reality and research

Purpose – The purpose of this paper is to present a review of the deception detection literature that arrives at a different conclusion from the one presented by King and Dunn. Specifically, the authors’ review shows that people can detect deception at significantly above chance accuracy in policing environments. A new paradigm for deception detection is also discussed.Design/methodology/approach – An extensive literature review was conducted.Findings – People can detect deception at levels that exceed chance in a variety of police‐related environments when an ecological approach to detecting deception is adopted.Practical implications – The authors’ review suggests that it is time for deception detection training and manuals to move away from the demeanor‐based systems that are currently dominant and toward coherence and correspondence‐based systems.Originality/value – The paper presents a perspective that is different from the one advanced by King and Dunn. It also introduces the ecological detection of...

The Impact of Accusatory, Non-Accusatory, Bait, and False Evidence Questioning on Deception Detection

This research examines question effects in deception detection. A first set of participants (N = 104) were given the opportunity to cheat to obtain a cash prize, and were then interviewed with accusatory, non-accusatory, bait, or false evidence questioning. A second set of participants (N = 157) watched videotapes of the interviews and made honesty judgments. Finally, interviewee behaviors were coded for demeanor. Overall, accuracy was high (72% overall, 70% excluding confessions, and 62% excluding confessions and adjusting for base rate). The type of question set made little difference in truth bias, accuracy, or demeanor, but false evidence questioning yielded 80% confessions compared to 20% confessions with non-accusatory questioning. No false confessions were obtained.

A critical look at meta-analytic evidence for the cognitive approach to lie detection: A re-examination of Vrij, Fisher, and Blank (2017)

Vrij, Fisher, and Blank (2017) published the first meta-analysis assessing the effectiveness of Vrij’s (2015) new cognitive approach to lie detection. Their basic premise is that verbal andnonverbal deception cues canbe amplified bymaking lyingmore cognitively effortful. Cues can be amplified in threeways: (1) instilling additional cognitive load, (2) prompting additional information, and (3) employing unexpected questions. Amplified cues lead to more accurate deception detection. Vrij et al. (2017) reported that their cognitive approach showed a 15-point advantage over the traditional standard approach. They claimed 71% correct classification for their cognitive approach compared to 56% for the standard approach in head-to-head experimental comparisons. This essay provides a critical look at Vrij et al.’s (2017) claims, methods, and data. Vrij et al.’s meta-analysis averaged across two conceptually, empirically, and pragmatically different types of outcomes. As a result of conflated outcomes, evidence for the cognitive approach was exaggerated and important patterns of findings were hidden. Specifically, accuracy in the controls and reliability in assessment correlate negativelywith support for the cognitive approach.

Producing deception detection expertise

Purpose – To function effectively, police must separate lies from truth. Police, ideally, would be experts at this task, yet there is debate surrounding whether expertise in detecting deception is possible. Drawing upon literature outside of deception detection, the purpose of this paper is to explore whether subjects making deception judgments can improve their performance. Design/methodology/approach – The sample was 19 students from two graduate-level classes. Subjects viewed six sets of videos over eight weeks. The first five sets displayed individuals reporting whether they cheated on an exam. The sixth set displayed individuals reporting whether they had committed a (mock) robbery. After each video, subjects judged whether the videoed individual was truthful, and then the actual status was revealed. Findings – Subjects’ accuracy improved consistently over the first five sets of videos; from about 69 percent accuracy to about 89 percent accuracy. However, the accuracy for the sixth set of videos drop...