Rethinking Gun Control, Part 6: Spree Shootings
Throughout my attempts to statistically analyze gun violence (on the state level, nationally, and internationally), one criticism I received from multiple readers was that I wasn't looking at rates of so-called "spree" killings. 2012 was marked by an unusually high number of such mass murder stories, leading many--the liberal Mother Jones and left-wing pundit Rachel Maddow, among others--to conclude that spree killings are on the rise. In order to test this hypothesis, I set out over the course of the last few weeks in search of raw data to support or contradict it.
|Most articles like this include pictures of mass shooters, but since I refuse to do that, here's a picture of Christian Bale mourning the victims of the Aurora shooting|
First, you have to define what you mean by "spree killing." The relevant FBI definitions are for "mass murder," which is "a number of murders (four or more) occurring during the same incident, with no distinctive time period between the murders," and "spree murder," which is "two or more murders committed by an offender or offenders, without a cooling-off period." However, the FBI, after discussing the matter "at great length," concluded that "spree murder" was too arbitrary to include as a separate statistical category.
Either way, finding specific FBI statistics online for mass murders involving a firearm proved too difficult, so I was forced, when consulting FBI data, to restrict my data collection to homicides with multiple victims, and I could only collect that data from the years 2008-2012. The following chart is based on that data:
For the purposes of finding spree shooting trends, this chart isn't terribly helpful. For one thing, it doesn't separate out homicides not caused by firearms, nor does it go back far enough in time. I also wasn't able to adjust for population, since reliable population statistics haven't yet been compiled for 2011 or 2012. Still, it does help illustrate that, in terms of people being murdered in groups, 2012 wasn't significantly worse than the previous four years. There may be a slight upward trend, but it would most likely disappear if you take population into account.
What am I to make, then, of graphs that appear to contradict the FBI data? For example, the aforementioned Mother Jones article includes this graph:
The article claims that researchers at Mother Jones conducted over five months worth of data collection--from multiple, frequently unnamed sources--to identify 62 mass shootings in the period between 1982 and 2012. The Mother Jones criteria define a mass shooting as any shooting that occurred in a public place in which at least four people (potentially including the shooter) were killed and only one shooter was involved, though there is wiggle room for Mother Jones to arbitrarily include "a handful" of other events. Though I won't claim to have sifted through all the data, I will point out that, in October 2002, Robert S. Flores, Jr. shot and killed three professors and himself at the University of Arizona, and yet, even though this incident fits the criteria, it is missing in the above graph.
The Mother Jones conclusions are still compelling, but they are also contradicted by this graph from James Alan Fox, a professor of criminology at Northwestern University and well-known "Dean of Death" for his exhaustive research into mass murder:
Fox claims to have assembled the data from official police reports to the FBI and concludes that "the facts say clearly that [there] has been no increase in mass killings, and certainly no epidemic." In Fox's contention, clustering of shooting events is coincidental and not reliant on any trend. He points out that there was a "flurry of postal shootings" in the 1980's, which the Mother Jones data all but ignores.
One must also be careful of statistical sleight-of-hand. For example, Rachel Maddow's contention that spree shootings are on the rise is based on a definition where at least 12 victims were killed. If you look at it that way, six out of the last twelve such instances of "mass bloodshed" have occurred in the last six years. However, if you change the cutoff point to 14 victims, as Paul Campos at Time points out, you could argue that spree killings were twice as frequent between 1966 and 1991 as they have been since (four to two).
In any event, regardless of how you define "spree shooting" or how you arrange the data, the fact remains that the chances of you or anyone you know being involved in such a violent and terrible crime are imperceptibly small. For instance, if you look at the 2009 data from the FBI, the second highest year for multiple victim homicides on my graph, the total number of homicides involving multiple victims is 1,428. This is liable to include all spree shootings, along with much else. The average total population for the United States in 2009 was 307,330,000, meaning that, if you were around in 2009, you had a roughly 0.0005% chance of being one of multiple homicide victims. Consider that, during the 2009-2010 swine flu pandemic, the CDC estimates that 12,000 Americans died, which we can be absolutely certain is at least an order of magnitude greater than the number of people who were killed in spree shootings during the same time period.
Still, for those 0.0005%, their friends, family, and colleagues, the grief is very real. Statistics have a way of sounding cold, and it is not my intention to downplay the horror that a spree killing brings with it. However, when making wide-reaching policy decisions, priorities matter. Is it worth risking our civil liberties to address a 0.0005% chance? I am certainly willing to entertain methods that can reduce that risk even more, but I am not willing to listen to the shrill voices telling me that they are certain spree killing is some kind of growing American epidemic that is threatening our way of life. So really--and I mean it this time--I think I should switch my focus to the policy side of the equation and leave the statistics behind.
-e. magill 11/12/2013