Shootings and Mass Murders Seem to Be Contagious

Data spanning decades shows how high-profile events can cause outbreaks of similar killings that mirror the spread of disease

Roses mark a window filled with bullet holes after a shooting spree in Denmark. Pacific Press/Corbis

School shootings and mass killings sometimes seem to be spreading across America like a disease. Now, a mathematical analysis of hundreds of past events suggests a disturbing explanation: such high-impact violence may actually be contagious.

Based on multiple data sets spanning events from 1998 to 2013, scientists tracked the timing of mass killings and shootings in the United States. They found that between 20 to 30 percent of mass killings—attacks that cause four or more deaths—seem to happen in bunches that mirror the outbreak of a contagious disease.

A few killers actively admit to being influenced by previous murderers, like several school shooters who referenced the 1999 Columbine High School massacre as a factor in their later acts. Sociologists have long suspected that such a link may exist in a considerable number of cases, but it's been difficult to tell for sure whether murderers were inspired by others—especially when many commit suicide. So Sherry Towers of Arizona State University and her colleagues set out to try to quantify the effect on a large scale for the first time.

No comprehensive federal database yet exists on mass killings—though Towers stresses that one is desperately needed—so her team mined other sources of information. The Brady Campaign to Prevent Gun Violence compiled data on 188 school shootings from 1998 to 2013, and on 477 mass shootings (three or more shooting victims but not necessarily deaths) from 2005 to 2013.

To examine mass killings (those with four or more deaths) the researchers used data compiled in a USA Today study that found 232 incidents between 2006 and 2013. The data set was based on reports to the FBI but also on records at local police departments, which often aren't sent to federal authorities. The USA Today report highlighted many omissions and inaccuracies in the reporting of such crimes, so Towers and her team only included data they could verify. 

When reviewed as a whole, the combined data revealed that mass killings and school shootings do appear to create a contagious period that lasts about two weeks, during which time similar crimes are more likely to occur, the team reports this week in PLOS ONE. Previous research suggests that other types of crime and violence may also be contagious. It appears that suicides, for example, can sometimes cause other vulnerable people to take their own lives.

“In many big cities they don't publish much information about suicides that might occur on a bridge, for instance, because it's been noticed that suicide can be contagious, so they don't give them a lot of coverage to try to help suppress that,” Towers notes.

The authors believe that nonstop news coverage of high-profile killings and the people who perpetrate them likely plays a role in any contagion effect that does exist. Details of these events more easily reach the relatively small number of people disturbed enough to consider committing more, the authors note. This idea is supported by the fact that contagion doesn't appear to exist for crimes that don't get as much national attention. The team's examination of mass shootings involving three or more victims but not necessarily any deaths found no evidence of contagion. 

“We found that those events were so common that they rarely made it past the local news,” Towers explains. “So we theorize that's why we found no statistical evidence for contagion in that data set. It's not a large number of people who would do these things, it's very few people, so the local coverage just doesn't reach a big enough pool of people. I think it may be that only the national coverage is able to reach enough people.” 

Media coverage is far from the only factor involved, Towers stresses. Clearly mental health issues play a major role. And according to the study data, access to guns does as well. “One other thing that popped out is that incidents are a lot higher in states that have a high prevalence of firearm ownership,” Towers says. “That appears to be playing a role. You get somebody who is emotionally or mentally distressed, then they also have access to weaponry, and it's a bad combination.”

Get the latest Science stories in your inbox.