An official website of the United States government
Objectives: Eyewitness research has adapted signal detection theory (SDT) to investigate eyewitness performance. SDT-based measures in yes/no tasks fit well for the measurement of eyewitness performance in show-ups, but not in lineups, because the application of the measures to eyewitness identifications neglects the role of fillers. In the present study, we introduce a SDT-based framework for eyewitness performance in lineups—Multi-d′ Model. Method: The Multi-d′ model provides multiple discriminability measures which can be used as parameters to investigate eyewitness performance. We apply the Multi-d′ model to issues in eyewitness research, such as the comparison of eyewitness discriminability between show-ups and lineups; the influence of lineup bias on eyewitness performance; filler selection methods (match-to-description vs. match-to-suspect); eyewitness confidence; and lineup presentation modes (simultaneous vs. sequential lineups). Results: The Multi-d′ model demonstrates that the discriminability of a guilty suspect from an innocent suspect is a function of discriminability involving fillers; and underscores that the decisions that eyewitnesses make in lineups can be regarded from two perspective—detection and identification. Conclusions: We propose that the Multi-d′ model is a useful tool to understand decisionmakers’ performance in a variety of compound decision tasks, as well as eyewitness identifications in lineups.
more »
« less
A multi-item signal detection theory model for eyewitness identification
Abstract How do witnesses make identification decisions when viewing a lineup?Understanding the witness decision-making process is essential for researchers to develop methods that can reduce mistaken identifications and improve lineup practices. Yet, the inclusion of fillers has posed a pivotal challenge to this task because the traditional signal detection theory is only applicable to binary decisions and cannot easily incorporate lineup fillers. This paper proposes a multi-item signal detection theory (mSDT) model to help understand the witness decision-making process. The mSDT model clarifies the importance of considering the joint distributions of suspect and filler signals. The model also visualizes the joint distributions in a multivariate decision space, which allows for the incorporation of all eyewitness responses, including suspect identifications, filler identifications, and rejections. The paper begins with a set of simple assumptions to develop the mSDT model and then explores alternative assumptions that can potentially accommodate more sophisticated considerations. The paper further discusses the implications of the mSDT model. With a mathematical modeling and visualization approach, the mSDT model provides a novel theoretical framework for understanding eyewitness identification decisions and addressing debates around eyewitness SDT and ROC applications.
more »
« less
- Award ID(s):
- 2017046
- PAR ID:
- 10629946
- Publisher / Repository:
- Springer Science + Business Media
- Date Published:
- Journal Name:
- Cognitive Research: Principles and Implications
- Volume:
- 10
- Issue:
- 1
- ISSN:
- 2365-7464
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
IntroductionDespite converging evidence that people more closely associate the construct of criminality with Black people who exhibit a more African facial phenotype than Black people who express a more European phenotype, eyewitness researchers have largely ignored phenotypic bias as a potential contributor to the racial disparities in the criminal legal system. If this form of phenotypic bias extends to eyewitness identification tasks, eyewitnesses may be more likely to identify Black suspects with an African rather than European phenotype, regardless of their guilt status. Further, in cases where the witness’s description of the perpetrator does not contain phenotypic information, phenotypic mismatch between the suspect and the other lineup members may bias identification decisions toward or against the suspect. If witnesses can use elements of the lineup construction to guide their identification decisions rather than relying on their recognition memory, then the lineup should be deemed unfair due to suspect bias. The current study also investigated lineup presentation method as a procedural safeguard, predicting that that when lineups were presented simultaneously, there would be a significant two-way interaction of phenotypic bias and lineup composition, with a larger simple main effect of phenotypic bias when lineups were suspect-biased (i.e., the fillers were a phenotypic mismatch to the suspect) than when all lineup members shared the same phenotype. We expected that this interaction would be significantly smaller or non-significant for sequential lineups. MethodsParticipants watched a mock crime video that contained a Black culprit with either a more African phenotype or a less African phenotype before attempting identifications from a photo array that contained a suspect whose phenotype always matched the culprit viewed in the video, but varied in culprit-presence, phenotypic match of the suspect and fillers, and presentation method. ResultsParticipants did not identify Black suspects with Afrocentric features more often than Black suspects with Eurocentric features. However, witnesses made more identifications of suspects when the fillers did not match the suspect’s phenotype compared to when all lineup members possessed similar phenotypic features. DiscussionIn sum, phenotypic bias did not influence our participant-witnesses’ identification decisions, nor interact with lineup composition and lineup presentation type to affect identifications of suspects, suggesting that phenotypic bias may be less influential in match-to-memory tasks than other types of legal decision-making (e.g., determining guilt and sentencing). However, the suggestiveness created by failing to match fillers’ phenotypes to the suspect’s phenotype can be avoided with proper attention to fair lineup construction.more » « less
-
SDT-based measures in yes/no tasks fit well for the measurement of eyewitness performance in a show-up (a detection task), but not in a lineup (a compound task of detection plus identification). We introduce a new SDT-based framework for eyewitness performance in lineups (Multi-d′ model). The Multi-d′ model demonstrates that the discriminability of a guilty suspect from an innocent suspect is a function of discriminability involving fillers and the differential filler siphoning effect; and that eyewitnesses’ discrimination in lineups can be assessed at two levels—detection and identification levels. We apply the Multi-d′ model to issues in eyewitness research.more » « less
-
Abstract AI assistance is readily available to humans in a variety of decision-making applications. In order to fully understand the efficacy of such joint decision-making, it is important to first understand the human’s reliance on AI. However, there is a disconnect between how joint decision-making is studied and how it is practiced in the real world. More often than not, researchers ask humans to provide independent decisions before they are shown AI assistance. This is done to make explicit the influence of AI assistance on the human’s decision. We develop a cognitive model that allows us to infer thelatentreliance strategy of humans on AI assistance without asking the human to make an independent decision. We validate the model’s predictions through two behavioral experiments. The first experiment follows aconcurrentparadigm where humans are shown AI assistance alongside the decision problem. The second experiment follows asequentialparadigm where humans provide an independent judgment on a decision problem before AI assistance is made available. The model’s predicted reliance strategies closely track the strategies employed by humans in the two experimental paradigms. Our model provides a principled way to infer reliance on AI-assistance and may be used to expand the scope of investigation on human-AI collaboration.more » « less
-
Abstract Early‐age hydration of cement is enhanced by slightly soluble mineral additives (ie, fillers, such as quartz and limestone). However, few studies have attempted to systematically compare the effects of different fillers on cementitious hydration rates, and none have quantified such effects using fillers with comparable, size‐classified particle size distributions (PSDs). This study examines the influence of size‐classified fillers [ie, limestone (CaCO3), quartz (SiO2), corundum (Al2O3), and rutile (TiO2)] on early‐age hydration kinetics of tricalcium silicate (C3S) using a combination of experimental methods, while also employing a modified phase boundary and nucleation and growth model. In prior studies, wherein fillers with broad PSDs were used, it has been reported that between quartz and limestone, the latter is a superior filler due to its ability to partake in anion‐exchange reactions with C‐S‐H. Contrary to prior investigations, this study shows that when size‐classified andarea matchedfillers are used—which, essentially, eliminate degrees of freedom associated with surface area and agglomeration of filler particulates—the filler effect of quartz is broadly similar to that of limestone as well as rutile. Results also show that unlike quartz, limestone, and rutile—which enhance C3S hydration kinetics—corundum suppresses hydration of C3S during the first several hours after mixing. Such deceleration in C3S hydration kinetics is attributed to the adsorption of aluminate anions—released from corundum's dissolution—onto anhydrous particulates’ surfaces, which impedes both the dissolution of C3S and heterogeneous nucleation of C‐S‐H.more » « less
