Jump to comment:

• Evaluation of firearm regulations
  David Hemenway
  Published on: 13 May 2016

• Published on: 13 May 2016

  Evaluation of firearm regulations

  • David Hemenway, Professor

  Dear Editor,

  In a short article, Rosengart et al. (“An evaluation of state firearm regulations and homicide and suicide death rates”) attempt to evaluate the effect of 5 different state regulations on four different outcomes (firearm suicide, suicide and firearm homicide, homicide). There is much that is good about this article, but a problem, perhaps due to space constraints, is that the authors do not discuss s...

  Show More

  Dear Editor,

  In a short article, Rosengart et al. (“An evaluation of state firearm regulations and homicide and suicide death rates”) attempt to evaluate the effect of 5 different state regulations on four different outcomes (firearm suicide, suicide and firearm homicide, homicide). There is much that is good about this article, but a problem, perhaps due to space constraints, is that the authors do not discuss some important limitations or caveats for many of their analyses. These limitations substantially reduce the likelihood that the authors will detect any effect of four of the laws (minimum age for purchase; for possession; one-gun-per-month; junk gun ban).

  For example, they evaluate the effect of laws raising to (or lowering from) 21 the minimum legal age for (a) handgun purchase and (b) handgun possession. Changing the minimum legal age from 18 to 21 or 21 to 18 mainly affects 18-20 year olds. Ideally, the evaluation should focus on the effects on this age group, with comparisons to other age groups. But the only age comparisons provided are for ages 0-19, and 20 and over. This evaluation doubly dilutes the possibility of finding any effect--by including many age groups with the 18-20 year olds, and then by cutting the 18-20 group in two. The authors are in effect seeking to find a measurable impact of the law on the overall rate of (say) firearm suicides. They are very unlikely to find such an effect, even if it exists. In 2002, for example, 18-20 year old firearm suicides accounted for only 3.7% of all firearm suicides. If the law could have had an immediate and incredible effect of reducing 18-20 year old firearm suicides by 25%, the effect on overall firearm suicides would have been less than 1%, not readily detectable by the evaluation. [For homicides, since the effect of the law should primarily be observed on perpetrators age 18-20, an evaluation with the power to detect an effect would probably need to use 18-20 year old perpetration data rather than victimization data].

  The authors also try to evaluate the effect of one-gun-per month laws. The authors’ essentially are comparing what happened pre and post in a state like Virginia which passed a one-gun-per-month law, to what happened in other states like New York, which did not pass such a law. A problem with their approach is that one purpose of these laws is to reduce gun running across state lines, from states with permissive gun laws (e.g. Virginia) to those with more restrictive laws (e.g. New York). A one-gun- per-month law in Virginia may help reduce gun availability and homicide in New York. But in the authors’ calculations, these lives saved would be counted against the law’s effectiveness.

  I believe the authors try to do too much. They focus on the evaluation of concealed carry laws, but then add, with little modification or discussion, evaluations of four additional laws. But each of these laws deserves its own careful and nuanced evaluation. I do not have a strong priors concerning whether any of these four laws had any effect, but I don’t believe the paper by Rosengart et al. provides a proper evaluation. Given their methodology, I am not at all surprised that the authors find little evidence “that any of the (four) laws evaluated were associated with a significant reduction in either firearm homicide or firearm suicide rates.”

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top