Making Sense of Mass-Shooting Statistics

James Alan Fox vs. Mother Jones

Last month I wrote a post here casting doubt on the claim that the number of mass shootings in the United States has been increasing. One of the criminologists I cited was James Alan Fox, an expert on mass murders who teaches at Northeastern University in Boston. Another was Grant Duwe, who works for the Minnesota Department of Corrections. I also noted in passing a Mother Jones feature that came to a different conclusion than I did, and I linked to Michael Siegel's criticisms of the Mother Jones piece.

Now Fox has written his own critique of the Mother Jones article, arguing that the magazine's decisions about which crimes to count and which to ignore are "hard to defend." Duwe, meanwhile, has presented a count of mass public shootings. Looking back at 2012, he reports that there were more of these incidents than usual last year but also that this was preceded by a 12-year decline. (Both Fox and Duwe are dealing in raw numbers, not incidents or casualties per capita, though Fox does address the country's population growth in the comment thread below his essay.)

As I noted in my original post, clusters of these crimes occasionally occur close together, a phenomenon Fox attributes to a mix of copycatting and coincidence. It's too soon to tell whether Duwe's figure for 2012 represents a new trend or another cluster, but Fox offers a good reason to expect the latter:

What is abundantly clear from the full array of mass shootings, besides the lack of any trend upward or downward, is the largely random variability in the annual counts. There have been several points in time when journalists and other people have speculated about a possible epidemic in response to a flurry of high profile shootings. Yet these speculations have always proven to be incorrect when subsequent years reveal more moderate levels.

The year 1991, for example, saw a man kill 23 people at a cafeteria in Killeen Tex., and a disgruntled graduate student murder five at the University of Iowa, along with other sensationalized incidents. The surge in mass killings was so frightening that a rumor spread around the nation that there would be a mass murder at a college in the Northeast on Halloween. Fortunately, October 31 came and went without anything close to a massacre taking place.

Two years later, in 1993, the nation was shaken by a series of workplace shootings, which encouraged a number of syndicated talk shows to air special programs about "ticking time bombs at the office." Despite the sudden spike in workplace homicide, the incidence of workplace murders actually declined throughout the 1990s.

In his 1999 book Random Violence, which I recommend highly, the sociologist Joel Best points out that "criminologists usually doubt claims about crime waves. Crime waves, they say, are really waves in media attention: they occur because the media, for whatever reason, fix upon some sort of crime, and publicize it." Genuine spikes in crime do occur, of course, but the press has a habit of spotting patterns that aren't there. That's worth remembering in all kinds of contexts.