Are Mass Public Shootings Seasonal? (Corrected)

***CORRECTION: A FACEBOOK READER noticed something in the Mother Jones data I presented recently that I had missed. Beginning in 2013, MJ changed their definition of a mass public shooting from 4 or more victims to 3 or more victims in 2013 (see more below), but did not retroactively update their database. Although not deceptive (they said plainly they were doing this, I simply missed it), this is methodologically problematic. So I eliminated those cases, which reduces the total number in the database from 114 incidents to 95, and re-did the chart here.***

I was sitting at the park tonight, watching my wife play tennis and minding my own business, when Rob Pincus texted to ask me if I had any source for data on the seasonality of mass shooting activity. I.e., mass shootings by month.

I did not, but I was intrigued enough by the idea to do a little work when I got home tonight. The fruit of that labor is below. Important notes and interpretive points follow the chart.

ANALYSIS AND INTERPRETATION

Average number of mass public shootings by season (corrected):

• SPRING (March, April, May) = 7.0
• SUMMER (June, July, August) = 7.7
• FALL (September, October, November) = 8.3
• WINTER (December, January, February) = 8.7

I did not statistically analyze the difference in number of shootings by season, but both eyeballing the chart above and looking at the average number of shootings by season suggest there is not a major seasonal effect on mass public shootings.

To the extent there is a seasonal effect, it is not what I expected. I had expected summer to have the highest number of mass public shootings, as I thought that homicides in general were higher in the summer time. (Turns out that may not be the case, either; it’s complex.) In these data, winter (corrected) is the mass public killing season, especially relative to spring.

DATA NOTES

I use data on mass public shootings compiled by Mother Jones magazine. Anyone can download, scrutinize, and analyze this data. Regarding the data: There is no commonly agreed upon or “official government” definition of a mass shooting. So, to be clear, Mother Jones has a particular definition of a mass shooting:

Our research focused on indiscriminate rampages in public places resulting in four or more victims killed by the attacker. We exclude shootings stemming from more conventionally motivated crimes such as armed robbery or gang violence.

I have no problem with this working definition. Everyday murders that claim a large number of victims or domestic violence murders (typically murder-suicides) in which someone wipes out a family do not generally make the entire public anxious in the same way that mass public shootings do.

***CORRECTION: As noted above, the working definition used by Mother Jones changed in the middle of their dataset. Mother Jones was not being sneaky in doing this, as hey make clear on their site:

In January 2013, a mandate for federal investigation of mass shootings authorized by President Barack Obama lowered that baseline to three or more victims killed. Accordingly, we include attacks dating from January 2013 in which three or more victims were killed. (Any analysis of the frequency of mass shootings should account for this.)

So, it was my fault for not making clear that they changed their inclusion criterion for 2013 forward. That said, I don’t think methodologically it is a good idea to make a change like this without retroactively updating the entire database using the new definition.***

* * *

FINAL NOTE: I didn’t bother to do a Google search on this, so if some other blogger or scholar out there has already addressed this issue, my apologies for not citing your work. This was just meant to be a quick and dirty look at the issue in response to Pincus’s question.