With American Sniper topping Box Office sales for weeks on end, it’s no surprise that the trial of the man accused of murdering the real-life protagonist, Navy SEAL Chris Kyle, is also making headlines. And though a jury was selected on Monday, defense attorneys were understandably worried that, with all the pre-trial publicity, they wouldn’t be able to find impartial jurors.

Research suggests it’s a legitimate concern. In a 2007 paper in the journal Applied Cognitive Psychology, a team of psychologists at the University of South Florida— Christina Ruva, Cathy McEvoy, and Judith Becker Bryant—designed an experiment to simulate the effects of pre-trial publicity. They recruited 588 university students, whose ages ranged from 18 to 52, and had them watch the tape of a real trial of a man accused of murdering his wife. (He pled “not guilty,” claiming she had accidentally fired a gun into her head while he tried to talk her out of killing herself.)

At the first session, participants were divided into two groups; one group was asked to read newspaper articles that painted the defendant in a negative light or cast doubt on his story. (For instance, one article claimed that the defendant had a bad temper; another said that his wife didn’t know how to use a gun.) Meanwhile, the other group read articles with no relevance to the trial. When the participants were brought back to the lab several days later for the second session, they were shown a 30-minute video of the trial and asked whether they believed the man was guilty, and—if they did—how long a prison sentence they would recommend. (Some participants deliberated in a group, like a jury; others had to give verdicts without consulting anyone else.)

As Ruva and her colleagues expected, they found that negative pre-trial publicity can have “an extremely biasing effect on juror decision-making.” The group that had been exposed to the negative articles was significantly more likely to find the man guilty; they also recommended, on average, a significantly longer sentence. The exposed and non-exposed groups recommended sentences of 40.7 years and 37.7 years, respectively.

The participants were also given a set of statements about the trial, and told to indicate what the source was—whether they believed they’d learned it from the pre-trial publicity or during the trial itself. If participants had been exposed to the unflattering articles, they misremembered the source of about 15 percent of the statements. Ruva speculates that in real life, the error rate is likely to be much higher: The publicity wouldn’t be limited to a single occasion several days before the trial, but would come in a constant stream, in a variety of media.

Of course, even if pre-trial publicity were somehow abolished, members of juries—representative as they are of the society at large—would still be biased. Other research has shown that juries dole out harsher penalties to defendants who are non-whitenon-citizensobese, or unattractive.