Fool me once, fool me twice

: the relationship between statement consistency and veracity across repeated recalls

Abstract

The overarching aim of the current thesis is to examine the relationship between veracity and consistency across multiple recalls provided by a single individual in the context of repeated investigative interviews. As a form of communication, deception is a dyadic process which can be examined at two points: sending and receiving. The content of an interviewee’s statement can be examined to determine whether it is consistent over multiple presentations, and a veracity assessor’s judgment-making process can be examined to determine the influence of their personal beliefs about the relationship between consistency and deception on their perception of the interviewee’s statements. 
Chapters II and III examine consistency at the point of sending. In Experiment 1, I examined how truth tellers’ and liars’ consistency differed across two phases of an interview as a result of strategic responses to manipulated interview techniques. In Experiment 2, two types of free recall written interviews were compared from the first interview. After a one week delay, truth tellers and liars participated in an oral interview. Across both experiments I found little support that objective measuring of could be used to indicate veracity. Chapters IV and V of this thesis examine consistency from the receiver’s perspective. In Experiment 3, participants were asked to rate a series of interview transcripts for consistency, based upon objective coding criteria for consistency, and to separately make a holistic consistency judgement. In Experiment 4, participants analysed four interview transcripts and reported which factors they found to be indicative of (in)consistencies. Based on this explanation, I created the Layperson’s Understanding of Consistency Key (LUCK) coding scheme, which was then compared to a coding scheme built around the experimental definition of consistency. I found that holistic consistency assessments could predict veracity where objective consistency coding could not (Expt. 3 & 4), suggesting that the lay assessment 3 of consistency relies on features not captured by researcher-generated coding schemes (Expt4). 
This set of experiments is the first to highlight the importance of differentiating between the examination of consistency as a sending or receiving pattern. Across the programme of research, I found little support for the objective measuring of consistency to infer veracity. However, holistic consistency judgements were found to be relatively accurate at differentiating between truthful and deceptive statements. In the discussion of the results, I consider the applied implications of using consistency to infer veracity, and suggest avenues for future research to decipher the mechanisms that inform the more reliable holistic consistency judgements.

Date of Award	Sept 2020
Original language	English
Awarding Institution	University of Portsmouth
Supervisor	Aldert Vrij (Supervisor), Lucy Akehurst (Supervisor) & Lorraine Hope (Supervisor)