Eyewitness Identification Accuracy and Response Latency: The Unruly 10 –12-Second Rule Nathan Weber and Neil Brewer Flinders University Gary L. Wells Iowa State University Carolyn Semmler and Amber Keast Flinders University Data are reported from 3,213 research eyewitnesses confirming that accurate eyewitness identifications from lineups are made faster than are inaccurate identifications. However, consistent with predictions from the recognition and search literatures, the authors did not find support for the “10 –12-s rule” in which lineup identifications faster than 10 –12 s maximally discriminate between accurate and inaccurate identifications (D. Dunning & S. Perretta, 2002). Instead, the time frame that proved most discriminating was highly variable across experiments, ranging from 5 s to 29 s, and the maximally discriminating time was often unimpressive in its ability to sort accurate from inaccurate identifications. The authors suggest several factors that are likely to moderate the 10 –12-s rule. Since the mid-1990s, postconviction use of forensic DNA evi- dence has led researchers to believe that mistaken eyewitness identification is the primary cause of the conviction of innocent people (Wells et al., 1998). Years before forensic DNA could be used to confirm this eyewitness misidentification problem, labo- ratory research by psychologists suggested that eyewitness identi- fication evidence had two troublesome properties: (a) mistaken identifications can be quite high under some conditions, and (b) once mistaken identifications occur, they are not easily distin- guished from accurate identifications. In the current article, we are concerned with the second of these problems: What is there to distinguish between accurate and inaccurate identifications? Spe- cifically, we examined the “10 –12-s rule” which proposes that lineup identifications faster than 10 –12 s maximally discriminate between accurate and inaccurate identifications (Dunning & Per- retta, 2002). Research directed at finding “markers,” or assessment variables as they have been named by Sporer (1993), that might help crime investigators, judges, and juries distinguish accurate and mistaken eyewitness identifications has focused primarily on eyewitness confidence. Although early research, which relied on the point- biserial correlation to probe the confidence–accuracy relation (e.g., Bothwell, Deffenbacher, & Brigham, 1987; Wells & Murray, 1984) suggested that eyewitness identification confidence had little or no relation to eyewitness identification accuracy, research in recent years has been more promising. Using confidence– accuracy calibration, rather than the standard correlation statistic, a number of studies have reported impressive confidence–accuracy relationships with both eyewitness identification (Brewer, Keast, & Rishworth, 2002; Brewer & Wells, 2004; Juslin, Olsson, & Winman, 1996) and face recognition paradigms (Olsson, Juslin, & Winman, 1998; Weber & Brewer, 2003, 2004). We are interested here in another promising marker of eyewit- ness identification accuracy that has emerged in recent years, namely identification response latency. Response latency 1 refers to the amount of time that an eyewitness takes to make and indicate his or her identification decision from a lineup. Research has shown that there is a negative statistical relation between response latency and accuracy in eyewitness identifications from lineups. Witnesses who make accurate identifications make their decisions faster than witnesses who make inaccurate identifications (Dun- ning & Stern, 1994; Smith, Lindsay, & Pryke, 2000; Smith, Lindsay, Pryke, & Dysart, 2001; Sporer, 1992, 1993, 1994). Sporer (1992, 1993) suggested that lineup decisions are the result of a process of sequential comparison of each lineup member with an image of the offender in memory. When the target is compared with the image (i.e., a correct match), the large number of features in common allows a very fast decision. In contrast, a lineup foil (i.e., an innocent filler) will not have as many features in common with the image in memory and will, therefore, be matched more slowly. Thus, the negative relationship between accuracy and response latency arises. This negative relation between response latency and identification accuracy appears to hold only for choos- 1 We follow convention from the processing speed literature in referring to the time that elapses from presentation of the lineup to when the participant indicates his or her choice as “response latency,” rather than decision time or latency, because it incorporates both decisional and motor components. Nathan Weber, Neil Brewer, Carolyn Semmler, and Amber Keast, School of Psychology, Flinders University, Adelaide, South Australia; Gary L. Wells, Department of Psychology, Iowa State University. This research was supported by Grant A00104516 from the Australian Research Council to Neil Brewer and Gary L. Wells. Correspondence concerning this article should be addressed to Neil Brewer, School of Psychology, Flinders University, GPO Box 2100, Adelaide, South Australia 5001, Australia. E-mail: neil.brewer@flinders.edu.au Journal of Experimental Psychology: Applied Copyright 2004 by the American Psychological Association 2004, Vol. 10, No. 3, 139 –147 1076-898X/04/$12.00 DOI: 10.1037/1076-898X.10.3.139 139 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.