Thursday Recap

Sunday Meeting Highlights

The Curious Asymmetric Confidence-Accuracy Relationship for Positive Identifications and Lineup Rejections

Speaker: John Wixted, University of California, San Diego

Summary by Xueqing Chen, University of Bristol, UK

John Wixted, a Distinguished Professor in the Department of Psychology at UCSD, has dedicated his career to unraveling the complexities of human memory. With extensive research into the basic mechanisms of memory and its forensic applications, Wixted's work has significant implications for addressing wrongful convictions tied to eyewitness errors—one of the leading causes of wrongful accusations. His innovative approach integrates psychological theory with legal practice to reduce the prevalence of eyewitness misidentification.

Eyewitness misidentification has been recognized as a primary cause of wrongful convictions, with two main threats undermining its reliability: forgetting and contamination. Memory accuracy is at its highest immediately after an event; however, as time passes, correct memories fade, and incorrect memories can gain strength. In the real world, law enforcement may not test eyewitness memory until much after the critical first moments have passed. By the time of trial, these memories may become unreliable. Despite this, Wixted's research highlights that initial identifications made with high confidence are often 95% accurate, suggesting that misidentifications at this stage are relatively rare.

A surprising trend emerges when examining the initial tests of eyewitness memory: many individuals confidently identified at trial were initially rejected by the same witnesses. This early evidence of innocence is crucial, yet these cases often lack DNA evidence for exoneration. Recent examples, such as the exoneration of Miguel Solorio by the Northern California Innocence Project, illustrate this point. Solorio, wrongly convicted of murder, was rejected by four witnesses in the initial lineup but was later misidentified by two of the same witnesses at trial due to contaminated memory.

In his latest research, Wixted employs model-fitting techniques to examine why confidence in lineup rejections remains weak. Contrary to the traditional belief that confidence in a lineup rejection is based on the average strength of memory signals from the lineup, his findings suggest that the strongest memory signal still governs confidence, even when a lineup is rejected. This challenges existing assumptions about how witnesses process and reject lineup participants. His findings highlight the need for an enhanced understanding of memory dynamics during eyewitness identifications.

Wixted's insightful talk highlights the urgent need to focus research efforts on the initial outcomes of eyewitness tests. The weak correlation between the confidence eyewitnesses express in their identifications and the actual accuracy of those lineup rejections demands a thorough re-evaluation of existing methodologies. Wixted advocates for an integrated approach, where scientific research informs and shapes legal protocols, enhancing the safeguards against wrongful convictions and ensuring that justice is served based on the most reliable evidence available.

Schemas of Science: Does Understanding How Science Works Impact Science Learning?

Speakers: Rian E. Drexler, University of California, San Diego, Celeste Pilegard, University of California, San Diego

Summary by Hannah Mechtenberg, University of Connecticut, USA

American’s trust in science has never fully recovered to its pre-pandemic levels. This lack of trust presents a significant issue for current and future policymaking that needs to rely on scientific insight to guide us towards a safer and more equitable society. Further, as covered in another talk session on Sunday morning by outgoing Psychonomic Society Governing Board Chair Dr. Stephan Lewandowsky, American congressmembers are moving away from using evidence-based language (e.g., facts, expert, accurate, etc.) and towards intuition-biased language (e.g., doubt, feeling, etc.). So, what can experts do to try and improve science literacy and general scientific knowledge?

Rian Drexler, a PhD student at the University of California San Diego, is working towards answering that exact question. Drexler began by breaking science into two components: processes and products. Science processes are the “how” of science—that it is based on evidence, humans do it, and it is inherently iterative. Together, this is known in the literature as the “nature of science” or a person’s epistemic beliefs about science. Meanwhile, the products of science are what we get: vaccines, knowledge about how evolution works, neuroanatomical models, etc. Drexler asked if learning about science products (typically what we teach about science) could be improved if learners also had specific instruction on the nature of science to provide a schema to fit the novel scientific information into.

Caption Image Above: Data from a newly released report by the Pew Research Center into Americans’ trust in scientists. While the majority of respondents believe scientists are intelligent, only 45% think that scientists are effective communicators. Looks like we can all brush up on our communicative abilities.

In an empirical study with 211 undergraduate students, Drexler tested if specific instruction in the nature of science would boost comprehension in a scientific subject highly familiar to us cognitive psychologists: selective attention. Participants were randomly sorted into a control group (no nature of science instruction) and the experimental group, where they received instruction on the nature of science. Both groups then watched a short lecture on selective attention. Afterward, participants completed a test of the selective attention content, and those in the experimental group were also tested on their knowledge of the nature of science.

First—the good news. Participants who watched the nature of science lecture demonstrated significantly better comprehension of scientific processes than those without, suggesting they were engaged in learning that important content. Unfortunately, that learning did not lead to better comprehension of the selective attention literature above the control group.

Caption Image to the Left: Drexler introducing schemas and how they might help learn novel scientific information.

While these results were not expected, and may feel disappointing, this work is a critical step towards addressing issues in scientific literature and trust in science writ large. While we wait for Drexler’s next study, perhaps we can all bolster our science communication efforts with a little sprinkling of the “how” of science. Maybe a glimpse behind the curtain will help everyone trust each other a bit more.

Caption Image to the Left: Data from Drexler’s study showing that participants who received information about the nature of science had higher comprehension for the nature of science than participants who did not have that instruction. 

Protective Effects of Multilingualism on Cognitive Aging Among Older Adults within the Context of Socioeconomic Status and Urbanicity

Speakers: Iris Strangmann, Columbia University Irving Medical Center, Sarah Petrosyan, University of Southern California, Erik Meijer, University of Southern California, Emma Nichols, University of Southern California, Shrikanth Narayanan, University of Southern California, Jinkook Lee, University of Southern California, Miguel Arce Rentería, Columbia University Irving Medical Center

Summary by Melinh Lai, University of Chicago, USA

Iris Strangmann presented an overview of her work investigating the degree to which cognitive abilities are better preserved in older adults who know more than one language. Theoretically, the ability to reason and problem-solve in multiple languages, and thus flexibly navigate different systems of grammar, words, and sounds, might help a person sharpen their cognitive skills over their lifetime, including as an older adult. But the circumstances under which multilingualism benefits cognitive abilities are unclear, including when these benefits are applicable and who they are most likely to help. Adding to the difficulty of answering these questions is the fact that the ability to learn multiple languages often depends on factors of the environment, like socioeconomic status (SES) or where a person lives (e.g., in a city versus in the countryside). To add even more complication, these factors can influence cognitive abilities independently of a person’s level of multilingualism. As an example, Strangmann cited a previous project from her team that found broad patterns of cognitive benefits driven by the level of knowledge of multiple languages in people from middle or upper SES groups, but no such benefit of multilingualism for people of low SES.

The goal of the project that Strangmann presented was thus first to investigate whether multilingualism did indeed have a positive association with older adults’ cognitive skills. In addition, she and her colleagues were also interested in whether details about individual or socio-environmental contexts modulated the relationship between multilingualism and cognitive aging.

The authors thus conducted a longitudinal aging study in India as part of the initial wave of their long-term project, Diagnostic Assessment of Dementia (LASI-DAD), which surveyed hundreds of older adults to gather a rich dataset matching demographic details with measures of cognitive ability at multiple points in time. Strangmann and her team classified their participants according to multiple factors, including whether they were multilingual, lived in an urban or rural environment, and came from a low, medium, or high SES. They also collected measures of participants’ memory skills, executive functioning, language abilities, and visuospatial processing through multiple cognitive skills tests. These social factors and cognitive measures were modeled against each other, with participant age and sex as covariates, to investigate the cognitive benefits of multilingualism potentially interacting with either their living environment (urban or rural) and their SES (low, medium, high).

Looking across the entirety of their dataset, the researchers found that being multilingual did indeed seem to be associated with good performance on all four cognitive skills. Moreover, generally speaking, the degree of this beneficial influence of multilingualism seemed to be further modulated by SES and living environment, such that participants from low SES had little to no benefit from multilingualism, and neither did participants living in rural areas. Instead, the strength of multilingualism’s protective qualities on cognition was greater for participants in medium SES and/or living in a city and greater still for participants from high SES and/or urban environments.

Thus, in answer to their research questions, the researchers found evidence that knowing multiple languages could indeed be a protective factor of a person’s mind as they age. However, other factors such as socioeconomic status or urbanicity also influence the rate of cognitive aging. While these factors could contribute independent effects than multilingualism, they are also tightly linked to the quality of being multilingual. After all, learning another language is often the product of schooling, which is unfortunately more available to higher SES groups than to lower ones. The results of this longitudinal study suggest that SES and urbanicity do indeed modulate the extent to which multilingualism protects an aging mind, with multilingual benefits being absent in lower SES levels and weaker for participants in rural environments. 

Thursday Recap