Search for a command to run...
ABSTRACTResearch on confession has usually focused on showing that it is significantly associated with individual, crime-related, and situational/contextual variables and is both a static event and a dichotomous indicator of interview success. Recent work, however, suggests that investigative interviews are a dynamic process in which interrogation strategies change over time. Using a Game Theory perspective, this study looks at the impact of behaviours of both players (interviewer and suspect) on the production of investigation-relevant information (IRI). The sub-objective is to demonstrate the usefulness of applying Game Theory to the study of investigative interviews by considering time and interaction between players as an integrative part of the analysis. Videotaped interviews related to online child sexual exploitation (n = 130) were analysed and the different behaviours of suspects and interviewers were analysed to determine if they involved (1) rapport building/active denial, (2) collaboration, (3) confrontation, (4) emotion/response, and (5) elicitation of information related to the case. Results showed that information relevant to the investigation is often provided shortly after a suspect has offered additional information or given responses that meet emotional needs (e.g. justifications). The interviewer's use of available evidence increases the likelihood that additional information will be provided, while the ability to build a rapport with the suspect is effective in the longer term, even if a positive effect is not immediately observed. Using a dynamic process approach in analysing investigative interviews provides a starting point for the creation of practical guidelines to help practitioners increase suspect collaboration during investigative interviews.KEYWORDS: Investigative interviewsconfessiondisclosurepolice interrogationGame Theorydynamic interaction Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 All the interviews were conducted in French.2 The research assistant was first assigned five interviews which was then compared with the first author's codification on these five interviews. Throughout the codification of these interviews, the research assistant regularly met with the first author to ensure proper and valid coding. After this training phase, since the level of error was low, the research assistant proceeds to code by herself almost all of the remaining interviews included in the research project (88%). As such, the interrater reliability was not calculated.3 As explained earlier, in the Canadian context only the interviewer and the suspect are usually allowed in the interview room. Some of the behavior included in the Kelly et al (Citation2013) taxonomy has also been adapted. For example, the taxonomy included interviewer behaviors such as bluffing or lying about evidence, but these techniques are illegal in Canada and were not observed in the sample.4 The complete list of behavior included in each category and its frequency is shown in Appendix 1.Additional informationFundingSSHRC - CRSH : 435-2019-0734.