A recent study has shed light on the factors that significantly impact eyewitness identification accuracy, with key findings emphasizing the critical role of distance, lighting conditions, and facial masking. The study, conducted with over a thousand participants observing live targets (“criminals”), provides valuable insights into the challenges faced by eyewitnesses and highlights the implications for criminal investigations. The new findings have been published in Psychology, Crime & Law.
The ability to recognize a person’s face is a specialized skill that surpasses recognition of other objects. This skill is crucial in various contexts, especially in legal proceedings where eyewitnesses play a pivotal role in providing evidence. However, despite its importance, eyewitness identifications are susceptible to errors, which can have serious consequences in criminal cases.
Recent research has primarily focused how law enforcement should conduct and construct a lineup, but less attention has been given to variables like facial masking, viewing distance, lighting, and eyewitness age that play a role when a victim or witness observes the crime. This study aimed to address these gaps in understanding and provide a more comprehensive view of how these factors affect eyewitness identification accuracy.
“No earlier research has investigated the combined effects of increased distance, decreased lighting, and facial masking on eyewitness identification accuracy using live mock criminals. This lack of earlier research is in itself surprising given the high likelihood of a crime taking place during such conditions,” said study author Thomas Nyman, an assistant professor at New York University Shanghai.
“Moreover, we wanted to know if there are visual thresholds for eyewitness accuracy. That is, if you see someone in twilight conditions, is there a distance threshold where it becomes impossible to create a clear memory of the criminal’s face. The witness choosing someone from a later lineup would be completely random and could of course also lead to an innocent suspect being identified.”
“Second, this line of research, that is, investigating the effects of distance and lighting on eyewitness accuracy, was the main focus of my PhD (freely available in English here) and continuing this research by investigating the effects of facial masking, was a natural progression.”
In the study, participants observed live targets (i.e. suspected criminals) and answered questions about the person’s gender, height, age, weight, and distance from them. Following this, participants were shown an eight-person line-up and asked to identify the “criminal” if present or reject the lineup if the “criminal” was absent. They also rated the certainty of their choice.
The study involved 1,425 participants, and after data cleaning, 4,456 responses were analyzed. The participants’ ages ranged from 5 to 90 years and were categorized into four groups: young children (5-11), older children (12-17), young adults (18-44), and older adults (45-90).
The researchers employed a mixed design with 48 different conditions. They manipulated factors such as distance (5, 12.5, or 20 meters), lighting (suboptimal: ~2 lux and optimal: ~300 lux), and facial masking (no masking, sunglasses, hood, and sunglasses and hood).
The study confirmed that greater distance between the eyewitness and the “criminal” had a significant negative impact on identification accuracy. For example, the accuracy for young adults dropped from 96% at 5 meters to 42% at 20 meters.
Reduced lighting, such as twilight or low-light conditions, also had a detrimental effect on identification accuracy. However, increased distance had a more pronounced negative impact than lighting.
Facial masking, particularly internal masking (altering internal facial features), had a substantial negative impact on identification accuracy. Sunglasses, in particular, had a more adverse effect compared to other masking methods.
“Perhaps the most surprising result was how quickly accuracy dropped when our mock criminals wore sunglasses – even as 5 meters there was a dramatic drop in accuracy between not wearing and wearing sunglasses,” Nyman told PsyPost. “The decrease in accuracy was much more substantial compared to the only slight decrease in accuracy when comparing the effects of daylight and twilight conditions without facial masking.”
Overall, the study highlighted that when “criminals” were observed under suboptimal conditions (e.g., low lighting and significant distance), the negative effects of facial masking were less prominent, as the overall identification accuracy was severely compromised.
“A main take-home from this research is the realization that eyewitnesses can make rather substantial errors depending on the visibility during the time they see a criminal and these results can be used in a court of law to assess the reliability of an eyewitness identification,” Nyman explained.
The study’s findings provide valuable guidance for practitioners in assessing the evidentiary value of eyewitness identifications in real-life criminal cases. Understanding these factors can help improve the reliability of eyewitness testimony and enhance the fairness of the criminal justice system.
“Naturally, our results are in many ways rather intuitive and a reader may ask why is this kind of research even necessary,” Nyman told PsyPost. “A way to respond to that question is to imagine that you are, hypothetically speaking, apprehended by the police despite the fact that you are innocent of the crime they have apprehended you for. Imagine also that the main source of evidence they have is an eyewitness who saw the perpetrator in poor visual conditions and that the eyewitness identifies you (mistakenly) from a line-up.”
“A mistake can happen because there are usually 6-8 images in a line-up and an eyewitness with a bad memory may select your image simply due to bad luck. If this is the case, and if the police rely on this evidence, then you could use our article to argue in court that the eyewitness testimony is not reliable and that it should not be used as evidence against you. In other words, it is direct and substantial support for the claim that certain contexts make eyewitness identifications unreliable as evidence.”
Despite the insights provided by this study, there are some limitations to consider. Firstly, the study did not replicate a real-life crime scenario, potentially limiting its generalizability. Secondly, participants were aware of the study’s purpose, which may not reflect the real-world unpredictability of eyewitness situations.
“Overall, there are many avenues to investigate further,” Nyman said. “Such as, investigating the impact of seeing multiple perpetrators in varying lighting conditions or at varying distances with or without facial masking on identification accuracy. Nevertheless, one thing I would like to see is a systematic investigation into how often these factors (e.g., increased distance, decreased lighting, various facial masking) are present in real-life criminal events.”
“We do have some statistics, but at the same time, these factors are often either not recorded or, if recorded, not taken into consideration by police and courts. If recorded, it is also often the case that it is unclear how these factors impact eyewitness reliability – this is where our research comes in and we hope that it will help the police and courts.”
“This was the second largest eyewitness study in the world to have been conducted with live mock criminals (our earlier study was the largest – see here),” Nyman added. “Moreover, this was a citizen science project that we conducted in collaboration between The Heureka Science Centre (Finland) and Åbo Akademi University (Finland). Participants joined in the experiment during their visit to the Science Centre and we designed it to be an exciting eyewitness game. So, this was a very unique and fun way of engaging the general public in the experiment – this also means that our sample was varied and this strengthens our findings.”
The study, “The masked villain: the effects of facial masking, distance, lighting, and eyewitness age on eyewitness identification accuracy“, was authored by Thomas J. Nyman, Julia Korkman, James Michael Lampinen, Jan Antfolk, and Pekka Santtila.
