The Effect of Channels of Communication on Accuracy in Detecting Deception in High-Stakes situations

Abstract

Much of the past research into deception detection has utilised low-stakes lies as stimulus, with globally poor results in accuracy levels. The present research used real-life recordings of high-stakes lies to investigate a between-subjects model of four different channels of communication: Audiovisual; Visual Only; Audio Only; Transcript Only. The dependent variable was the accuracy score obtained in each channel of communication in detecting deception. Considering available research results, it was hypothesised that the Audio Only group would score significantly higher than the Visual Only group, the Audiovisual group would score significantly higher than the Transcript Only group, and that the Transcript Only group would score significantly higher than participants in the Visual Only group. The lack of research into the channel of communication of Transcript Only provided further rationale for the present study. Due to the high-stakes nature of stimuli materials it was hypothesised that all participants would score higher than chance. Each participant group (N=20) observed 20 clips of people making public pleas for information about a missing or murdered relative. Half of the clips included people involved in the crime (attempting to deceive the public) and the other half were innocent (truthful, and not attempting to deceive the public). Scores ranged between 50.8% accuracy (audio visual) and 56.5% accuracy (visual only). There was no statistically significant difference between mean scores, F(3,76)=.30, p=.826, η²=.01. T-tests were conducted to test accuracy levels within each group. Accuracy levels were not significantly above chance. Suggestions for further research are discussed.