Skip navigation

Examining Gun Control

Recent events involving gun violence—including the tragedy at Sandy Hook Elementary School, Vice President’s Biden’s gun task force, pending state and federal  gun control legislation and the ongoing public debate on guns and violence—led to an amazing discussion in my Scientific Controversies class a couple weeks ago on gun control.  This class was typical of my approach, in that it entailed both discussion of a technically  dense scientific paper, as well as a discussion of the legal, political, economic and policy implications of that science.

In that class, we worked through Kellerman’s scientific paper on guns and violence, Gun ownership as a risk factor for homicide in the home, including  several cogent critiques of it.  I am not going to reproduce here our entire class discussion, but rather highlight one key technical issue, and its broader policy implications.  In this study, Kellerman compares households where a homicide had happened, to nearby similar households  where no homicide occurred. He finds relatively more guns in the latter as compared to the former. This seems simple enough. But a serious problem lurks.

Technically, Kellerman found a correlation between homicide and guns: those households where homicides occurred had relatively more guns, as compared to homes without homicides.  Critically however, we don’t know what caused what. We do know (because Kellerman tells us) that these households differed in ways other than gun ownership. For example there was relatively more alcoholism in the homicide households.

Why did the households with the homicides have more guns? Perhaps the guns caused the homicides (i.e., Guns—> Homicide), implying that these homicides would not have happened if the guns were not there, and implying that controlling guns would reduce homicide. Alternatively, perhaps the residents of those households where homicides did occur felt more unsafe, and purchased guns for their own protection (i.e., Social Environment—>Gun purchase & Social environment—> Homicide). In this situation, controlling for the presence of guns would actually increase homicide. Based on Kellerman’s data, we simply do not know.

Exactly the same problem arises when we compare states with and without gun control—urban states and large cities tend to have stricter gun control and also higher incidence of homicide, and rural states have less strict gun control and lower incidence of homicide. For example, Vermont has extraordinarily permissive gun laws (i.e. no permit needed for concealed carry, with few exceptions), and a low incidence of homicide. What causes what?

Critically, the issues in Kellerman’s paper arise in many other contexts, when we attempt to compare the effectiveness of alternative policies, be they policies on gun control, welfare, economic development, or environmental protection. In theory, federalism gives the fifty states freedom to implement alternative policies, so that we can see what works and what doesn’t. But when we compare states with different policies, we run into the exact same problem that Kellerman ran into.

Likely, all of this is unsatisfying. But I think it is realistic and makes for a fascinating classroom discussion.

Craig Pease
Professor of Science and Law


Post a Comment

You must be logged in to post a comment.
%d bloggers like this: