The language of accurate and inaccurate eyewitnesses

Abstract

What is the best method for collecting an eyewitness’s level of confidence? In an attempt to answer this question, we used the machine-learning methodology applied by Dobbins and Kantner (2019) to investigate the natural language of accurate and inaccurate eyewitnesses. This method followed a “bag of words” approach which treated each individual word in an eyewitness confidence statement as a potentially predictive feature of identification accuracy. In contrast to previous investigations (e.g., Behrman & Richards, 2005), this method provided a much more fine-grained analysis of eyewitness verbal confidence statements. We find that verbal confidence statements contain much diagnostic information. Even more noteworthy, we find that verbal confidence statements contain unique diagnostic information that is not already captured by traditional indicators of identification accuracy such as numeric confidence ratings. However, the diagnostic value of a verbal confidence statement depends in part on the face recognition ability of the eyewitness: strong face recognizers provide more diagnostic verbal confidence statements than weak face recognizers. Together, these results shed important theoretical light on the relationship between language and memory. Moreover, the machine-learning methodology used in this study may help actors in the criminal justice system accurately interpret an eyewitness's verbal confidence statement.

The language of accurate and inaccurate eyewitnesses.

“…While there is much information about the value that confidence holds when a numeric rating scale is used, less is known about the diagnostic value of verbal confidence statements (Behrman & Richards, 2005;Cash & Lane, 2017;2021;Dobolyi & Dodson, 2018;Dodson & Dobolyi, 2015;Grabman, Dobolyi, Berelovich, & Dodson, 2019;Klobuchar, Steblay, & Caliguiri, 2006;Mansour, 2020). Wixted and Wells (2017) highlighted this as a priority for future research, about which they asked: Should a confidence statement be taken in the witness's own words (as in Klobuchar et al, 2006), or should confidence be recorded using an explicit 3-point rating scale (as in Wixted et al, 2016) -or should a 100-point scale be used?…”