Correct identification decisions: How to obtain them and how to judge them
False eyewitness identifications as well as erroneous judgments of identification decisions in court have been shown to contribute to wrongful convictions in many criminal cases (Garrett, 2008, 2011, 2012). The present dissertation reports three experiments, which were aimed to investigate a new system variable (Wells, 1978) to increase ... identification accuracy as well as to analyze fact finders´ judgmental processes when evaluating the accuracy of an identification decision. In Experiment 1, potentially beneficial effects of re-reading one´s own person description prior to the identification task were examined (Cutler, Penrod, O´Rourke, & Martens, 1986; Sporer, 2007). Moreover, description effects on identification accuracy were investigated under more realistic retention intervals, expecting a verbal facilitation effect instead of verbal overshadowing (Schooler & Engstler-Schooler, 1990). Participants watched a video film (1) without describing the perpetrator (no description control group), (2) with describing the perpetrator or (3) with describing and re-reading their own description prior to the identification task. Two days or five weeks later all participants were asked to identify the perpetrator in a target-absent or target-present lineup. As predicted from a context reinstatement framework, results revealed higher identification accuracy after re-reading compared to the control group. Moreover, there was a tendency for a verbal facilitation effect with the likelihood of a correct identification decision being almost three times higher when the perpetrator was described compared to the control condition.In Experiment 2, the Brunswikian lens model (Brunswik, 1956, 1965) was applied to the evaluation of eyewitness identifications. To explain observers´ judgment accuracy when evaluating the accuracy of an identification decision it was examined (1) which cues observers use to evaluate an identification decision ( subjective utilities ), (2) how they interpret and weight these cues, and (3) if these cues as perceived by observers are indeed related to identification accuracy ("ecological validities"). Study 1 presented participant-observers with literal transcripts of 48 choosers´ identification decisions, whereas Study 2 used the original videotapes. A think-aloud method was applied to make discriminating cues more salient to observers, which was compared to retrospective reasoning protocols. Both studies demonstrated that observers used the investigated cues as indicators of identification accuracy independently of type of decision protocol. However, only when think-aloud protocols were evaluated cues discriminated between correct and incorrect identifications resulting in a high correspondence between subjective utilities and ecological validities. Advantages of think-aloud methods and videotapes to increase fact finders´ judgment accuracy when evaluating identification decisions are discussed.In Experiment 3, persuasive effects of different peripheral measures (i.e., speech style and attributed witness traits) and ratings of different person and event description qualities on observers´ judgments were investigated. Results demonstrated that observers used these cues to make their judgments (e.g., several description qualities and perceived witness confidence). However, almost none of these measures have been shown to be diagnostic of identification accuracy. Thus, the use of invalid peripheral cues contributes to explain observers´ low judgment accuracy. To conclude, this dissertation suggests re-reading one´s own person descriptions as a promising approach to increase identification accuracy. As it is common police practice to ask eyewitnesses for a description of the perpetrator, re-reading one´s description is an easily applicable system variable that does not require any additional procedures, training, or resources. In contrast to former research (Alogna et al., 2014; Schooler & Engstler-Schooler, 1990), describing a perpetrator does not seem to impair identification performance under more realistic conditions.Considering fact finders´ judgmental processes when evaluating the accuracy of an identification decision, the present studies demonstrated that videotaped or transcribed identification statements indeed contain some valid indicators of identification accuracy that are perceivable and usable by observers. Especially the application of videotaped think-aloud protocols seems to be fruitful to make valid cues more salient to observers. However, future researchers are encouraged to test and optimize these instructions for the evaluation of identification decisions. Moreover, the Brunswikian lens model framework offers an appropriate method to contrast relationships between empirically valid and intuitively used cues. Thus, to increase fact finders´ judgment accuracy, the model further allows recommendations for an appropriate weighting of information contained in identification protocols.