Computer animations used in court colored by bias, researchers say
A courtroom jury views a computer animation of a vehicle accident or heinous crime. Does it help bring a conviction or acquittal? With no clear standards for animations that re-create incidents, the verdict is still out, and, for now, it may depend on which side created the simulation, researchers say.
In a study of 117 undergraduate students, psychologists discovered that movements in a sequence of events, as well as the duration portrayed, in animations such as re-creations of a crime can double an already troubling hindsight bias.
Hindsight bias has been linked to the traditional use of text and diagrams to re-create crimes, and is known to interfere with the ability to make fair decisions because it often leads jurists to exaggerate the predictability of past events.
"Many lawyers assume that computer video animations help clarify the evidence, and, therefore, help jury decisions to be fairer and more closely grounded in the facts," said Neal J. Roese, a professor of psychology at the University of Illinois at Urbana-Champaign. "Our findings indicate instead that a computer animation introduces its own additional bias, making people more punitive and more likely to hand out harsh penalties."
Roese and colleagues suggest that by viewing a computer-animated re-creation of an event, a persons confidence is heightened -- but not necessarily accurately. An animation, they say, provides movement as reconstructed by a prosecution, plaintiff or defense witness to reconfirm or heighten a jurists hindsight feeling that "I knew it all along."
In the study, published in the April issue of the journal Psychological Science, some students viewed computer animations of highway incidents prepared for real court cases by Eleventh Hour Animation of Skokie, Ill. The idea was to compare judgments made in foresight, where an outcome is not known, with hindsight, where the outcome is known. A control group viewed text-plus-diagram re-creations.
One animation, 19 seconds long, showed a car following an 18-wheeler on a two-lane highway. The car attempted to pass but collided with a truck coming from the opposite direction. An 11-second depiction showed a semi-trailer avoiding a slow-moving vehicle that was turning into its path on a two-lane highway; the truck collided with a bus coming from the opposite direction. Both films showed the events from birds-eye views.
Participants were told in advance that they would see cases in which accidents may have occurred. Some participants viewed the entire re-creations with the accidents shown, whereas others saw depictions that were stopped before the accidents occurred.
Participants seeing the outcome then were told to disregard their knowledge of it and put themselves in the shoes of those who had watched clips that did not show results. All participants estimated the likelihoods of the various outcomes, from no accident at all to the serious accidents that, in fact, did occur. The hindsight, widely observed in past research, was found again.
"Participants who see how an accident happens have a very difficult time disregarding this knowledge and cannot place themselves in the shoes of naïve observers who did not see the accident," Roese said. "This hindsight bias occurred regardless of whether participants watched the scenes in computer animations, or if they read text descriptions of the events."
The findings suggest that when it comes to issues of liability or negligence, judgments that hinge on assessing what defendants know at the time of their actions, rather than what came later, the use of computer animations might be an especially big problem.
"Some courts rule them as inadmissible, many do not," Roese said. "Supposedly an animation is based directly on the available evidence and the laws of physics, and when animations are allowed, it is under the assumption that the video accurately illustrates the event. But the truth is that all reconstructions of evidence contain inherent imprecision."
Animated reconstructions, the authors argue, illustrate far more information about an accident for after-accident observers to consider than the involved drivers ever had when their situation was still unfolding.
Co-authors with Roese were psychology doctoral students Florian Fessel and Amy Summerville; Justin Kruger, a former U. of I. psychology professor now with the Leonard Stern School of Management at New York University; and Michael Dilich of FORESIGHT Reconstruction Inc., a research and consulting firm in Northbrook, Ill., which specializes in vehicle-accident investigation, reconstruction, analysis and safety research.
Jim Barlow | EurekAlert!