Who’s a terrorist and who’s mentally ill? We looked at 10 years of news coverage to find out.

When a perpetrator is Muslim, there’s a 488 percent greater chance an attack will be called terrorism

Only 17 percent of the attacks we studied were perpetrated by Muslims — but of the 1,384 articles that mentioned terrorism, 77 percent were covering these attacks. Of course, other factors had an effect; if an attack was perpetrated by members of an identified terrorist group, then coverage was understandably more likely to mention terrorism. And when the perpetrator was known or suspected of having a mental illness, media coverage was less likely to refer to terrorism.

Even controlling for these factors, we found that the odds of an article using the term “terrorism” is 488 percent greater when the perpetrator is Muslim.

But when an attacker is white, it doesn’t increase the chance that media coverage will mention mental illness

To our surprise, we found no differences in references to mental illness based on whether the perpetrator was white.