Reducing wrongful convictions

  • November 19, 2012
Reducing wrongful convictions

Close your eyes and tell me what you are wearing today. What is the person next to you wearing? Our memory is capable of incredible things; however, it does not work like a tape recorder, neither when memorising an event, nor when attempting to recall it. In court, eyewitness evidence is regarded to be among the most incriminating type of evidence. However, it is not always accurate and can lead to miscarriages of justice. To date, over 330 wrongfully convicted people have been exonerated in the US by DNA testing, including 18 people who had been sentenced to death. On average, these people had served 13 years each in prison before they were exonerated and released. In the US over 75% of known wrongful convictions, many of them in rape cases, are at least in part due to mistaken eyewitness identification, making it the single greatest cause of wrongful convictions. Although witness memory is indisputably fallible, it is heavily relied on. And although a wealth of literature exists on eyewitness issues in general, relatively little is known about older witnesses. This is especially concerning as the proportion of the world population that is 60 years or older will rise from 11 per cent in 2011 to 22 per cent in 2050, reaching a total of two billion. Taken with the fact that older adults are found to live healthier and more actively to older ages, this rapid population ageing will likely lead to an increase in older adults being witnesses of crimes and generally being involved in the Criminal Justice System. Bearing in mind older adults’ declining sensory systems and memory, it is most important to get further insights into strategies that can enhance their witness performance. Person identification decisions are subject to mistakes in general. However, older adults are found to perform even worse in eyewitness situations than young adults, both regarding testimony and person identification. Attempts that have been made to improve witness performance in general have had limited success. Specific interviewing techniques, for instance, are often training-intense and time-consuming. An easy-to-apply and effective method to aid older eyewitnesses could, however, come through a consideration of the witness’ individual circadian rhythm. Circadian rhythms reflect roughly 24-hour cycles in biological and physiological processes, which impact, among other things, on attention processes and memory. These show regular peaks and declines across the day and people who perform at their optimal time of day, i.e. according to their circadian rhythm, are at their peak of cognitive functioning. Applied to eyewitness research, my PhD supervisor Katrin Mueller-Johnson was able to show that taking elderly witnesses’ optimal time of the day into account substantially improved the quality and quantity of their testimony. To pursue this further, I examined the impact of circadian rhythm on performance in an eyewitness identification task. Again, preliminary results were very promising. In a first study, the overall odds of performing correctly in a line-up were roughly seven times higher if participants were tested at their optimal time of day (80% correct) than at their non-optimal time (37% correct), an effect that was highly significant. This becomes even more interesting when looking only at line-ups in which the perpetrator was not present, which is the research equivalent to situations in which the police have arrested the wrong suspect. Here, the odds of correctly indicating that the perpetrator is not there were eight times higher if participants were tested at their optimal time (33% falsely identified someone) as opposed to their non-optimal time (80% of participants falsely identified someone from the line-up). Although these are just preliminary studies and there are, of course, limitations to this research, it nevertheless constitutes an important step in the investigation of possible improvements of the person identification process for older eyewitnesses. It shows the importance of studying what influences eyewitness testimony and reveals a promising approach to improving older adults’ person identification. In due course this could lead to more effective police investigations, better access to justice and more successful prosecutions and ultimately to a reduction in the number of wrongful convictions. *Katrin Pfeil [2012] is doing a PhD in Criminology. Picture credit: renjith krishnan and

Latest Blogs

How Cambridge Analytica influenced Nigeria’s elections

My research broadly tackles questions of technology, equity, and accountability within the Global South. I emphasise the West’s use of sub-Saharan Africa as a testing ground or laboratory for its nascent technologies before launching them in the West. This blog is an excerpt from my recent presentation at the 2022 Gates Day of Research, titled […]

Why algorithms are necessarily value-laden

Algorithmic decision-making systems applied in social contexts drape value-laden solutions in an illusory veil of objectivity. Machine learning plays an increasingly prominent role in mediating institutional decisions in everything from corporate hiring practices to criminal sentencing. This ongoing AI spring has invigorated discussions of the ethical dimensions of these techno-social arrangements. In particular, there is […]

Preparing for all scenarios in an unstable world

On March 9th, Mohamed A. El-Erian joined the Gates Cambridge community for a virtual fireside chat, where he discussed decision-making in conditions of uncertainty, the economic impact of the pandemic and relief efforts and the importance of diversity of thought and scenario planning. El-Erian is President of Queens’ College, Cambridge and Chief Economic Advisor of […]

How can the international community help Belarus?

Last Sunday represented a tipping point in the recent history of Belarus which has had an immediate effect on the lives of its citizens, including mine. Independent exit polls and observers representing the diplomatic community, verified by the crowdsourcing platform Golos, show that, had it been a fair and transparent election, the uninterrupted, 26-year-long reign […]