For the Record: Exploring variability in interpretations of police investigative interviews

Recent research (Haworth 2018) has demonstrated how investigativeinterview data are (unintentionally) distorted as they pass through the criminaljustice system, and the survey-based experiment we present here was designedto test our hypothesis that various aspects of the processing of police-suspectinterview data may have an impact on the quality of the official evidentialdocument produced. The quantitative and qualitative findings from thisexperiment shed light on, and provide a sound evidence base for this claim, ratherthan leaving it as an untested assumption. The experiment was designed totest each key aspect of the current process of the production of routine writtentranscripts of investigative interviews (ROTIs), focusing on the conversion fromspoken to written format, and the use of different transcription conventions, andit has enabled us to investigate which changes make the most difference in termsof the evidential quality of the end product, in order to effect a change in practicewhich will reduce or eliminate the effect of those changes. Our findings suggestthat when presented with a transcript of a police interview, we are significantlymore likely to (1) perceive the interviewee as anxious and unrelaxed, (2) interpretthe interviewee’s behaviour as being agitated, aggressive, defensive, and nervous,(3) determine that the interviewee is un-calm and uncooperative, and (4) deemthe interviewee’s version of events to be untrue, than we are if we listen to theoriginal audio recording. Moreover, subjects identified (a) consistency, (b) phraseand lexical choice, (c) emotion (crying/upset), (d) hesitation and/or pauses assignificant factors influencing participants’ perception and interpretation of theinterviewee and their story. This is particularly concerning as the latter twofeatures are not currently routinely included in police transcripts, and Haworth(2018) illustrates multiple ways in which transcripts might differ from the originalaudio recordings they are intended to replace, with respect to words and phrases,as well as general content. The findings presented in this paper provide a strongmotivation for further research into how we capture spoken interaction in legalcontexts, and they constitute something of a mandate for reform with respect tothe transcription of police interviews in the UK.