When reporting violent events to 911, callers who fail to express expected levels of emotion and urgency may trigger suspicion that they are participants in the crime they are reporting - potentially the first step toward a wrongful conviction, Cornell research finds.
In four studies involving roughly 1,800 civilians and 300 law enforcement officers who listened to real or simulated 911 calls, researchers identified five behaviors that might make a caller seem suspicious. Conveying strong emotion and urgency reduced suspicion, while callers perceived to be communicating poorly, guarded in providing information or trying to make a favorable impression drew more suspicion.
Those first impressions can be important, the scholars said, because once someone is a suspect, research has shown that investigators may become prone to confirmation bias - leading them to be more drawn to evidence that confirms, rather than contradicts, a caller's guilt. Emergency calls may appropriately be scrutinized for factual evidence, they said, but should not be used to infer guilt from how someone sounds - as promoted by some flawed police training.
"Having expectations for people's demeanor on 911 calls is dangerous," said Jessica Salerno, associate professor in the Department of Psychology and the College of Human Ecology, and associate member of the Cornell Law School faculty. "People communicate in very different ways due to cultural and personality differences, neurodivergence, disabilities, or just the intense stress of the situation. There is no 'one size fits all' for responses to traumatic events."
Salerno is the senior author of "From Caller to Suspect: Identifying Behaviors That Trigger Suspicion in 911 Calls," forthcoming in Law and Human Behavior, a peer-reviewed journal of the American Psychological Association. Co-authors include Megan Lawrence, a doctoral student in psychology, and first author Samantha Bean, a litigation consultant and member of Salerno's Social Psychology and Law Lab.
The paper is part of a broader initiative studying assumptions about suspicious behaviors and their validity at predicting guilt, work that includes development of a large database of 911 calls. The research is relevant, Salerno said, to a case in which a Texas man, Robert Roberson, was convicted and sentenced to death for murdering his infant daughter based on now-debunked "shaken baby evidence." Despite a team of medical experts and even the detective who originally arrested Roberson fighting for his exoneration, he faces execution in October. Hospital staff and police perceived Roberson as oddly unemotional when he reported the death - but Salerno said he has been diagnosed with severe autism, which means he lacked the ability to express emotion in ways people expect.
In two other cases supported by the Innocence Network, in which Salerno is involved as an expert witness, she said detectives became suspicious of women who called 911 to report infants with medical problems. The detectives had received training that purported to teach them how to recognize if someone committed the crime they are calling 911 about, based on their behavior on the call. Including checklists for "guilty" or "innocent" behaviors, the methodology is based on a retired deputy's graduate research. But Salerno said scientists have identified flaws in that research's methodology and at least six peer-reviewed studies have been unable to replicate its findings.
"There are going to be times when a detective's hunches about behavior are right, but unfortunately many times when they are not," Salerno said. "The problem lies in thinking we can accurately judge people's intent based on their behavior in these tragic moments."
In the new research, participants listened to real 911 calls reporting violent acts, including shootings, or scripted dialogue based on actual events, performed by male and female professional voice actors. In addition to emotion and urgency, study participants evaluated behavior related to callers' cognitive load (difficulty communicating), information management (withholding detail) and impression management.
Emotion was cited most frequently - by four out of five study participants - as increasing or decreasing their suspicion. Even when hearing the same 911 call script, violent events reported in a less emotional tone attracted significantly greater suspicion, Salerno said. Results overall were similar across study participants, except that law enforcement officers were more suspicious of male callers; otherwise, no gender bias was detected in laypeople.
Better understanding the factors that prompt early suspicion is key to preventing wrongful convictions that are very difficult to overturn, Salerno said.
"The best way to combat a lot of wrongful convictions is to try to prevent the wrong person from being suspected in the first place," Salerno said, "rather than hoping to fix it later."
In addition to Salerno, Lawrence and Bean, co-authors of the forthcoming research are Alia Wulff, a former postdoctoral researcher in Salerno's lab; Isabelle Reeder, graduate student at Fordham University; Nicholas Duran, associate professor at the University of Texas, Austin; and Saul Kassin, professor at John Jay College of Criminal Justice, part of the City University of New York.
The research was supported by the National Science Foundation and Arizona State University.
