Situational and personal determinants influencing eyewitness evidence : Meta-analytical syntheses

Abstract

This dissertation describes two meta-analyses that are concerned with factors affecting eyewitness testimony. Meta-analysis 1 investigated the so-called weapon focus effect (WFE). This describes the phenomenon that a person witnessing a criminal event where the perpetrator carried a weapon is later worse able to describe or identify the target person compared to eyewitnesses not being confronted with a weapon. A total of 23 research articles met the inclusion criteria. A significant WFE could be calculated for description accuracy of the target person, gu = 0.568, 95% CI [0.490, 0.647], k = 29, while results for identification performance failed to reach significance. A parallel analysis was conducted with studies that tested the notion that the unusualness of an object is responsible for this effect. Unusualness also affected person descriptions but not correct identifications. The effect on false identifications could not be investigated due to insufficient data within the studies included. With regard to extensive consequences an identification of an innocent person has, future research should necessarily focus on this.In meta-analysis 2, the effect of eyewitness´ age was examined, comparing elderly with young witnesses. Overall, 22 studies tested the hypothesis that older eyewitnesses are worse in identifying a target after observing a crime compared to younger adults. Significant effects in favor for the younger age groups were found for all dependent measures under investigation. Largest effects were found for foil identifications for both target-present (OR = 2.453 [1.794, 3.356]) and target-absent lineups (OR = 3.074 [2.310, 4.091]), indicating that older eyewitnesses were 2.5 to 3 times more likely to choose a wrong person from a lineup. Similar results were found for old-age faces. Cognitive impairments as well as a more liberal choosing behavior of older people are discussed when interpreting the results. To conclude, the present meta-analyses demonstrated that both the presence of a weapon as well as the age of someone observing a crime affect eyewitness performance. However, it should be noted that experiments are far from real cases as they do not reflect the stress levels real crime victims (or bystanders) are likely to experience. Hence, even larger effects as they were found in the present syntheses are expected in real cases. In both cases, factors under investigation were estimator variables and are therefore not modifiable by the police within relevant investigations. Nevertheless, policy makers and decision makers could be informed to arrive at better evidence-based decisions. Practical implications are also discussed. In the case of WFE, a lineup consisting of different weapons could be helpful as analyses demonstrated eyewitnesses to focus on central details like the weapon carried by the perpetrator. Information about the weapon could be additional evidence in police investigations. One possibility to reduce or even prevent a WFE could be to train people (e.g., employees of a bank or gas station) not to be affected as much by the phenomenon. Some studies demonstrated a higher description accuracy when witnesses were warned beforehand (e.g., Pickel, French, & Betts, 2003). For the age of witnesses, interventions should be investigated to make the choosing behavior of older persons more cautious as they are 2.3 times more likely to choose someone from a lineup in comparison to younger adults. Although, age differences in correct identifications are smaller than in foil identifications higher choosing rates increase both. Appropriate instructions as well as pre-identification procedures could therefore be helpful for the elderly to adopt a stricter decision criterion (Dysart & Lindsay, 2001; Wilcock & Bull, 2010). Finally, methodological aspects of meta-analyses, which became very popular in the last decades, are critically discussed.