For many years psychologists have conducted carefully balanced and controlled experiments in order to understand the processes involved with line-up identifications. While the results are useful, several researchers have questioned whether these experiments are applicable to real line-ups. In this study we compare the results from two specialist line-up suites with the results from ordinary police stations. Witnesses at the suites chose foils more often than witnesses at the police stations. However, characteristics of the crimes, witnesses and suspects at the suites differed from those conducted at the stations. Many of these characteristics are also associated with increased proportions of foils being chosen. For example, the specialist suites tended to accommodate more non-white suspects, more crimes of violence and, at the time of this research, their line-ups occurred longer after the event. Several of the methodological difficulties of using data from real line-ups are discussed and appropriate statistical techniques are introduced.
There is a greal deal of interest in eyewitness identifications within line-ups or parades; court cases are sometimes decided only on line-up identifications. While much laboratory-based research exists (see Wells, 1993, for a review), there is relatively little research using real witnesses (Yuille, 1993; exceptions include Fisher, Geiselman and Amador, 1989; Tollestrup, Turtle and Yuille, 1994; Yuille and Cutshall, 1986). Several researchers have questioned whether laboratory-based results will necessarily generalize to real eyewitness situations (Herrmann and Gruneberg, 1993; Yuille and Cutshall, 1989). These researchers argue that the emotion levels associated with viewing an actual crime, as well as realizing the importance of the decisions, cannot be mimicked in a laboratory simulation (see Christianson, 1992a, for a counter-argument). We do not wish to debate the relative merits of laboratory and field research here (see Wright, in press). As Tulving (1991) argues, both approaches should be used to explore any phenomenon.
The cooperation of the Metropolitan Police Service in allowing Anne
McDaid to cany out research into identification suites is gratefully acknowledged. The analysis for and preparation of this paper were conducted with support from a British Academy Fellowship awarded to Daniel Wright. We would like to thank several reviewers for useful comments on an earlier draft and Min Yang, of the Multilevel Models Project at the Institute of Education, for helping with the modelling.