Do They All Look Alike? An Exploration of Decision-Making Strategies in Cross-Race Facial Identifications

Canadian Journal of Behavioural Science, Apr 2004 by Smith, Steven M, Stinson, Veronica, Prosser, Matthew A

Variables that play a role in eyewitness testimony can be categorized as system or estimator variables (Wells, 1978). Lineup presentation and lineup fairness typify "system" variables, those that can be controlled by the legal system (Lindsay, Smith, & Pryke, 1999; Malpass & Lindsay, 1999). Estimator variables are measured after the identification to determine the likelihood of eyewitness error, such as the relationship between confidence and accuracy (Kassin, 1985; Lindsay, 1986; Sporer et al., 1995). Whereas system variables inform the process, estimator variables inform researchers how to assess identifications for idiosyncratic variance in accuracy.

Estimator or "postdicting" variables are of great importance to the legal process as they are frequently used in trials to assess the veracity of eyewitness testimony. Errors in systems that rely on human decisions for reliability are inevitable. Memory and facial recognition are complex processes that are vulnerable to error. Therefore, research that incrementally extends the ability to postdict the reliability of cyewitness memory must be actively pursued. Evidence that judgment strategy can act as an estimator variable has been presented by Lindsay and Bellinger (1999) and Smith et al. (2000). When mock witnesses are given a description of absolute and relative judgment strategies and are asked to indicate which style best typifies their decision process, those who report using a relative judgment strategy produced more false positive identifications than those who used an absolute strategy (Smith et al.). Clearly, an understanding of judgment strategies can incrementally improve the system's ability to predict identification error. However, research examining decision-making strategies has focused almost exclusively on samerace situations (Smith et al, 2001).

Overview of the Research

The goal of the current research was to explore if judgment strategies predict lineup identification accuracy in cross-race situations. For example, are persons engaged in an identification task with a same-race target more likely to use an absolute judgment strategy than those engaged in a cross-race identification task? Cross-race identification may induce a relative judgment strategy as the other-race member may be initially categorized as simply being of another race. The result could be biased encoding, where the eyewitness may be less inclined to process the information carefully. Then, when faced with a simultaneous lineup containing only other-race foils, the eyewitness may engage in a within-category comparison, choosing the person most resembling who they remember. This recall strategy would reflect a categorical encoding process. Alternatively, encountering a same-race individual may be more likely to provoke a more careful processing of the facial characteristics. Subsequently, eyewitnesses may be better able to engage in absolute judgment strategies when presented with the lineup. On that basis, we predicted that individuals completing a cross-race identification task would be more likely to use a relative judgment strategy than individuals engaging in a same-race identification task. Using both openended and closed-ended measures of decision-making, we tested this hypothesis with the hope of finding a viable explanation for the cross-race effect.