Developing interviewing techniques to enhance information elicitation and lie detection

Abstract

Effective and accurate lie-detection is advantageous in police investigations. One approach for maximising the available cues for investigators to assess is by increasing the information elicited from cooperative interviewees. This thesis contributes to the development of investigative interviewing techniques designed to amplify the verbal reporting differences between truth tellers and liars. Chapter 1 introduces the topic of lie-detection and emphasises the applied importance of developing valid tools to encourage interviewees to say more in forensic interviewing. Chapter 2 examines specially tailored Model Statements which have been emphasised to include additional spatial or temporal information. The objective of Model Statements is to enhance information elicitation from truthful interviewees by providing a highly detailed example of a statement emphasising either spatial or temporal information. Results of the Chapter 2 study show an information elicitation effect. Specifically, truth tellers were able to provide more information when they heard a Model Statement relative to liars, whose verbal output did not change. These differences in verbal output allow more interviewees to be accurately classified as honest or deceptive, using number of details as a dependent measure. Unlike previous research, this study (reported in chapter 2) found an interaction effect (veracity × interviewing condition) underpinned by a change in the verbal detail reported by truth tellers. That is, the Model Statements resulted in truth tellers providing more information than liars. A further examination using two new(more detailed) Model Statements revealed that the effects on participants were implicit(beyond conscious awareness). In other words, the Model Statements functioned by priming truthful participants to provide more spatial or temporal information depending on whether they listened to a spatially or temporally emphasised Model Statement. Chapter 3 builds upon the findings of the first study by reporting effects found in a study of Model Statement modality. A highly detailed Model Statement containing additional spatial and temporal information was presented to participants either as an audio recording or as a written statement. Similar results to those of Chapter 2 emerged: truth tellers (but not liars) reported more detailed statements in both Model Statement conditions, compared to the control condition. This information elicitation effect also resulted in enhanced lie-detection accuracy, again using number of details as a dependent measure. Chapter 4 comprises a critical analysis of the Model Statement literature to evaluate the efficacy of this tool in practice. Results suggest the Model Statement is not (yet) ready for practical usage, as its utility as a lie detection tool has not been fully established. I discuss the diverse range of existing Model Statement scripts, substantial diversity in the dependent measures examined in conjunction with this tool, and why some measures of verbal content may not be suitable as outcome measures. These findings suggested two possible PhD research trajectories: (i) further examination of the Model Statement's functionality, and/or (ii) development of a new approach to lie-detection built upon strategic differences between truth tellers and liars. I explored the second option in Chapter 5, investigating the effects of a new Asymmetric Information Management (AIM) technique designed to exploit strategic differences between truth tellers and liars. This tool was designed with a simple set of instructions to encourage truth tellers to provide more information, while encouraging liars to withhold information. The results show that the simple AIM instructions enhance information elicitation and lie detection accuracy. Both Model Statement and AIM techniques function by encouraging truthtellers to provide more additional information than they would normally, whilst having no significant impact on the verbal output of liars. That is, the Model Statement and AIM instructions encourage truth tellers (but not liars) to be more detailed in their recalls. The Model Statement provides an exemplar of how detailed an interviewee should be, which functions implicitly (see Porter et al., 2018). In contrast, the AIM instructions are explicit which is perhaps why they appear more effective at encouraging truth tellers to report more, while asymmetrically encouraging liars to withhold more information. The AIM study presented in Chapter 5 employs a between-subjects design. In Chapter 6, I investigated the use of the AIM technique using a within-subjects design in an online experimental context. This allowed me to control for individual differences in reporting style while testing the effectiveness of the AIM technique in a setting more amenable to practitioners. Findings suggest the AIM instructions encourage the elicitation of new information in this context, particularly for truth tellers. In sum, I have evaluated the Model Statement literature and extended the field of lie-detection via the creation of AIM – a new investigative interviewing technique. In my general discussion, presented in Chapter 7, I recommend future research focuses on retrieval methods that may be used in conjunction with the AIM technique, with the objective of supporting information elicitation in applied contexts. I also provide an overview of my doctoral study, including the theoretical and practical implications this work and directions for future research.

Date of Award	Jun 2021
Original language	English
Awarding Institution	University of Portsmouth
Supervisor	Ed Morrison (Supervisor), Alistair John Harvey (Supervisor) & Rachel Taylor (Supervisor)