Alan Kafka
Weston Observatory
Department of Earth and Environmental Sciences
Boston College
Dramatic events can sometimes make us think we are seeing patterns in data that aren't necessarily there.
My purpose in writing this is not to say that everything is OK, and is not to be insensitive to the pain and suffering experienced from horrific mass shootings. One senseless murder in any form is one too many. And personally, I think that there are too many guns and too much violence in our society, and I fully agree that this is the time to talk about guns, gun laws, and violence. But I do feel compelled to say something about the role of statistical analysis in helping us to discern whether patterns we think we are seeing, such as an increase in mass shootings, might or might not be there.
While we are having our very important discussions about the horror of mass shootings and what we can do about it, it's important to maintain an awareness of the power and limitations of statistical analysis of scientific data. Often when we think we know what's going on with some issue, statistical analysis teaches us that what we think we know is different from what we really do know, and that sometimes what we really know is not much at all.
Is there an increase in mass shootings or not? What I think these studies of mass shootings and megaquakes teach us is that you can think that you know something and yet a statistical analysis demonstrates that it ain't necessarily so.
I haven't analyzed the mass shootings data myself, but if it's anything like geophysics (and I suspect it is), my guess is that we really don't know if mass shootings are on the rise or not. I could easily envision statistical studies suggesting no increase in mass shootings presented at a research meeting where some experts agreed with the result, some didn't agree, some thought it was a great contribution, and some found problems with the methodology, etc. Granted that this is a geophysicist speculating on a social science result, but I think it's safe to say that it would be very unusual scientific result if such a mixed review wasn't the case. At this point I wonder if anybody really knows if mass shootings are on the rise or not.
This does not mean that the statistical analysis failed us. Instead, to me it shows that whatever signal there is of an increase that signal is either very subtle, or maybe not even there.
The discussion about what should and should not be done about guns and violence in the aftermath of tragic mass shootings is an important discussion. As responsible citizens we should do whatever we can to minimize violence in our society. And we should do whatever we can to build cities with seismic safety in mind so that we minimize the loss of life from earthquakes. Working towards creating safe schools, homes and workplaces does not require statistical proof of an increase.
Statistical analysis does help us to gain insight into whether what we think we are seeing is really there. And knowing what really is (and maybe isn't) really occurring is an important part of making good, informed decisions that make our world a safer place to live in. In the cases of mass shootings and mega-earthquakes the signal is apparently not strong enough to tell us whether there is or is not a pattern. But that's useful information too.
Even one mass shooting in a hundred years is one too many. Yet in the midst of our sorrow this week, perhaps it may be of some small comfort to know that the incidence of mass shootings (and megaquakes) are not rising as dramatically as it might look to us in the moment.
