What lies beneath: an examination of the underlying components of deception detection

Abstract

Little is known about the underlying components that are responsible for the processes involved with decision-making in deception detection. Throughout this dissertation, I explored previous (mis)conceptions about decision-making in deception and its detection. In Study 1, I used self-report and objective measures (i.e., coding of witness statements) to examine a long-held dichotomy in deception detection research – that those labelled “truthtellers” are exclusively honest and those labelled “lie-tellers” fabricate their entire accounts. My findings revealed that both groups incorporated truthful and deceptive elements into their accounts, in fact. Although it has been well-established that truth-tellers provide more information than lie-tellers (DePaulo et al., 2003; Vrij, 2008), I found that truth-tellers provided more accurate information than lie-tellers, whereas the groups did not differ in utterances of inaccurate details. Rather, lie-tellers omitted significantly more details than truth-tellers, particularly during the free recall phase of the interview. In Study 2, I examined whether deception detection is static, as has been implied by the field’s focus on post hoc decision-making (Shanks, 2017). Using a novel, dynamic approach to measuring deception detection, I found that decisions and biases changed over time. Observers in the control condition – who made decisions after viewing an interview, as is typical in the literature – were able to discern between lie-tellers and truth-tellers and they exhibited a truth-bias, replicating previous research. However, observers who rendered their decisions continuously held no biases and were insensitive to veracity. Lastly, in Study 3, I examined the effect of varying the focal element of the deception detection task (i.e., person, event, or detail) because the impact of question phrasing on veracity decisions was unexamined in previous research. There were no differences between question phrasing conditions, which implies that phrasing is not the source of variability in discrimination or accuracy results throughout the literature. Overall, this dissertation serves to validate and challenge long-standing notions within deception detection research.