Melanie Sauerland

Post-decision confidence, decision time, and self-reported decision processes as postdictors of identification accuracy

Abstract Post-decision confidence, decision time, and decision processes were evaluated concerning their usefulness for postdicting identification accuracy. One hundred and ninety-two participants witnessed a filmed theft and were tested with target-absent or target-present simultaneous lineups 1 week later. Post-decision confidence was positively associated with, and decision time negatively associated with choosers’ identification accuracy. For several measures of self-reported decision processes, the expected associations were non-significant. Using a decision rule including highly confident and fast participants led to more correct classifications than either variable alone. Significant associations between postdictors highlight the need to consider the different processes jointly as, in combination, they may be more useful in assessing identification decisions.

Person descriptions and person identifications: Verbal overshadowing or recognition criterion shift?

In order to test rival theoretical accounts of the verbal overshadowing effect, participants (N=144) either gave no description or provided written descriptions after viewing a crime on video. Half of the describers reread their target descriptions prior to attempting to identify the thief either in a target-absent or target-present lineup 1 week later. Rereaders showed a criterion shift, that is, they less frequently chose somebody from the lineup than the other two groups. Within rereaders, description accuracy and choosing were positively correlated whereas number of incorrect details reported were negatively correlated with choosing. For describers that did not reread their description, there was a significant positive correlation between description accuracy and identification accuracy, and a negative correlation between number of incorrect descriptors and identification accuracy. The results lend support to the decision criterion shift approach (Clare & Lewandowsky, 2004) and shed light on the decision processes underlying target description and identification.

Developmental trends in different types of spontaneous false memories: implications for the legal field

In an emerging area of memory research, it is becoming apparent that one particular type of false memory, called spontaneous false memory, follows a developmental trajectory that is the opposite of what is commonly assumed in false memory research - that is, spontaneous false memories are more likely to occur in adults than in children. The present study focused on developmental trends of different types of spontaneous false memories. Specifically, in the current study, 6-8 year-olds, 10-12 year-olds, and adults were presented with two methods to induce spontaneous false memories: (i) semantically related word lists that are commonly used to evoke spontaneous false memories [i.e, Deese-Roediger-McDermott (DRM) paradigm]; and (ii) a video in which related details were not shown but were presented during a recognition task. The results showed that children were more likely to form false memories than adults in the video false memory paradigm, whereas DRM false memories were more evident in adults than in children. Furthermore, we found that on a general level, DRM false memories were positively related to video spontaneous false memories. We explain that stimuli that contain obvious themes attenuate or even reverse developmental trends in spontaneous false memories.

Speaking is silver, writing is golden? The role of cognitive and social factors in written versus spoken witness accounts

Contradictory empirical findings and theoretical accounts exist that are in favor of either a written or a spoken superiority effect. In this article, we present two experiments that put the recall modality effect in the context of eyewitness reports to another test. More specifically, we investigated the role of cognitive and social factors in the effect. In both experiments, participants watched a videotaped staged crime and then gave spoken or written accounts of the event and the people involved. In Experiment 1, 135 participants were assigned to written, spoken-videotaped, spoken-distracted, or spoken-voice recorded conditions to test for the impact of cognitive demand and social factors in the form of interviewer presence. Experiment 2 (N = 124) tested the idea that instruction comprehensiveness differentially impacts recall performance in written versus spoken accounts. While there was no evidence for a spoken superiority effect, we found some support for a written superiority effect for description quantity, but not accuracy. Furthermore, any differences found in description quantity as a function of recall modality could be traced back to participants’ free reports. Following up with cued open-ended questions compensated for this effect, although at the expense of description accuracy. This suggests that current police practice of arbitrarily obtaining written or spoken accounts is mostly unproblematic.

Consistency across repeated eyewitness interviews: contrasting police detectives' beliefs with actual eyewitness performance.

In the legal system, inconsistencies in eyewitness accounts are often used to discredit witnesses’ credibility. This is at odds with research findings showing that witnesses frequently report reminiscent details (details previously unrecalled) at an accuracy rate that is nearly as high as for consistently recalled information. The present study sought to put the validity of beliefs about recall consistency to a test by directly comparing them with actual memory performance in two recall attempts. All participants watched a film of a staged theft. Subsequently, the memory group (N = 84) provided one statement immediately after the film (either with the Self-Administered Interview or free recall) and one after a one-week delay. The estimation group (N = 81) consisting of experienced police detectives estimated the recall performance of the memory group. The results showed that actual recall performance was consistently underestimated. Also, a sharp decline of memory performance between recall attempts was assumed by the estimation group whereas actual accuracy remained stable. While reminiscent details were almost as accurate as consistent details, they were estimated to be much less accurate than consistent information and as inaccurate as direct contradictions. The police detectives expressed a great concern that reminiscence was the result of suggestive external influences. In conclusion, it seems that experienced police detectives hold many implicit beliefs about recall consistency that do not correspond with actual recall performance. Recommendations for police trainings are provided. These aim at fostering a differentiated view on eyewitness performance and the inclusion of more comprehensive classes on human memory structure.

Memory impairment is not sufficient for choice blindness to occur.

Choice blindness refers to the phenomenon that people can be easily misled about the choices they made in the recent past. The aim of this study was to explore the cognitive mechanisms underlying choice blindness. Specifically, we tested whether memory impairment may account for choice blindness. A total of N = 88 participants provided sympathy ratings on 10-point scales for 20 female faces. Subsequently, participants motivated some of their ratings. However, on three trials, they were presented with sympathy ratings that deviated from their original ratings by three full scale points. On nearly 41% of the trials, participants failed to detect (i.e., were blind) the manipulation. After a short interval, participants were informed that some trials had been manipulated and were asked to recall their original ratings. Participants adopted the manipulated outcome in only 3% of the trials. Furthermore, the extent to which the original ratings were accurately remembered was not higher for detected as compared with non-detected trials. From a theoretical point of view our findings indicate that memory impairment does not fully account for blindness phenomena.

The effect of choice reversals on blindness for identification decisions

ABSTRACT The present study examined blindness for identification decisions from target-present (TP) and target-absent (TA) lineups using a field study methodology. Eighty pedestrians were exposed to a staged theft. Subsequently, they were asked to identify the thief and the victim from separate, simultaneous six-person lineups. Their identification decision concerning the thief lineup was manipulated such that participants’ selections were exchanged with a previously unidentified lineup member (choice exchange) and lineup rejections were turned into identifications (choice reversal). Participants were 7–10 times less likely to detect choice exchanges (66.7%) compared with choice reversals (11.2%). Furthermore, identification accuracy was not a prerequisite for detection. Thus, rejections and particularly selections made from both TP and TA lineups are susceptible to choice blindness. Finally, our study implies that for blindness in eyewitness identification decisions between-category changes (i.e. choice reversals) are easier to detect than within-category changes (i.e. choice exchanges).

Providing eyewitnesses with initial retrieval support: what works at immediate and subsequent recall?

The effect of retrieval support on eyewitness recall was investigated in two experiments. Based on the outshining hypothesis, Experiment 1 tested whether retrieval support enhances witness performance (compared to free recall) especially when witnessing conditions are suboptimal (e.g., because witnesses were distracted during the crime). Eighty-eight participants watched a videotaped crime with either full or divided attention and subsequently received retrieval support with the Self-Administered Interview© (SAI) or completed a free recall (FR). One week later (Time 2 – T2) all participants completed a second FR. Unexpectedly, retrieval support did not lead to better memory performance than FR under divided attention conditions, suggesting that retrieval support is not effective to overcome adverse effects of divided attention. Moreover, presence of retrieval support at Time 1 (T1) had no effect on memory performance at T2. Experiment 2 (N = 81) tested the hypothesis that these T2-results were due to a reporting issue undermining the memory-preserving effect of T1-retrieval support by manipulating retrieval support (SAI vs. FR) at T1 and T2. As expected, T1-retrieval support led to increased accuracy at T2. Thus, the beneficial value of T1-retrieval support seems greatest with high-quality T2-interviews. Interviewers should consider this when planning a subsequent interview.

The application of multiple lineups in a field study

Abstract The usefulness of multiple lineups was tested in a field experiment with nine different targets. Six hundred and forty-eight passers-by were asked for directions in the pedestrian zone of a university town. Subsequently, they were approached by a different person and asked to identify the target from portrait, body, and profile lineups. Additionally, participants were asked to identify a shopping bag that the target had carried. Two of the lineups were target-present, and two target-absent. Diagnosticity ratios (DRs) were computed for target/suspect choices, lineup rejections and foil choices. Compared to foil choices and rejections, target/suspect choices were most diagnostic of guilt. Here, the combination of lineups was superior over individual lineups. Lineup rejections showed some capability of establishing innocence, but with lower DRs than target/suspect choices. Here, combinations did not increase diagnosticity. The diagnosticity of multiple foil choices was acceptable for portrait face lineups but limited for all other lineups or combinations.

These two are different. Yes, they’re the same: Choice blindness for facial identity

We examined the manipulability of face identity judgements by combining a sorting task for unfamiliar faces with a standard test of choice blindness. In Experiment 1, 50 participants completed a sorting task and then justified grouping specific pairs of photos together or apart. On manipulated trials, the presented pairings were different from those the participants had actually produced. Detection rates for these identity manipulations were strikingly low (∼21%). Moreover, participants readily provided justifications for identity decisions that they had not made, typically referring to specific facial features. Experiment 2 was conducted along similar lines and confirmed that lower task difficulty and higher confidence in ones face identity judgements increase detection rates. We conclude that observers can easily be led to believe that they made identity judgements they did not make. As well as underscoring the fragility of unfamiliar face matching, our findings have implications for identity judgements in legal settings.