The malleability of eyewitness confidence: Co-witness and perseverance effects.

Abstract

A theft was staged 70 times for pairs of eyewitnesses (N = 140) who then made a photo-lineup identification. Witnesses then received 1 of 9 types of information regarding the alleged identification decision of their co-witness. Witnesses told that their co-witness identified the same person whom they had identified showed an increase in the confidence they expressed to a confederate police officer. Confidence deflation occurred among witnesses who thought their co-witness either identified another person or had stated that the thief was not in the lineup. Initial co-witness information was not mitigated by subsequent changes to that information. A second study showed videotapes of these witnesses' testimonies to observers (n = 378) whose credibility ratings of the testimony paralleled the witnesses' self-rated confidence. Eyewitness identification confidence is highly malleable after the identification has been made despite the fact that physical resemblance between the culprit and person identified has not changed. If an eyewitness says am absolutely confident that he is the guy I saw rob the liquor store, it is rather difficult for people to accept the idea that the witness could be wrong. Numerous studies have demonstrated a close relation between the confidence expressed by an eyewitness and people's propensities to accept that eyewitness's testimony as accurate (e.g., Brigham & Bothwell, 1983; Fox & Walters, 1986; Lindsay, Wells, & Rumpel, 1981; Wells & Leippe, 1981; Wells, Ferguson, & Lindsay, 1981; Wells, Lindsay, & Ferguson, 1979; Yarmey & Jones, 1983). It is probably quite rational to place more trust in the validity of someone's judgment when that person makes a statement with high confidence rather than with little confidence. In the case of eyewitness identification testimony, however, the observed empirical relation between accuracy and confidence is surprisingly weak. A meta-analysis by Bothwell, Deffenbacher, and Brigham (1987) indicates that eyewitness confidence accounts for about 6% of the variance in eyewitness identification accuracy. Hence, it is not uncommon in research to find eyewitnesses making false identifications with high self-rated confidence that they are correct (or, conversely, eyewitnesses being quite unsure and yet accurate in their identifications).