Showing papers in "Journal on firearms and public policy in 2009"

How not to study the effect of gun levels on violence rates

Abstract: Most research on the effects of rates of gun ownership on violence crime rates has little of a persuasive nature to say on the subject, because it is afflicted by the same simple methodological problems that have characterized this field of inquiry for decades: 1) the failure to properly model the possible two-way relationship between gun levels with the effect of gun levels on violence rates; 2) the use of invalid measures of gun levels; and 3) the failure to control for a substantial number of (or any) possible confounding factors, i.e. factors that influence violent crime rates but are also associated with gun levels. A recent article by Cook and Ludwig (Journal of Public Economics 2006; 90(1/2): 379-391. [SafetyLit note: Incorrectly cited by Kleck as "The cost of guns". The correct title is, "The social costs of gun ownership"]) is used to illustrate these problems, since it suffers from all three flaws. A review of prior research demonstrates how consequential these problems are, showing that the findings interpreted as indicating a violence-increasing effect of gun levels are entirely confined to research that did little or nothing to solve these problems. Finally, it is shown that it is perfectly possible to avoid the problems, and that very different findings are obtained when one does so. The handful of studies that have seriously addressed all three problems consistently find no significant positive effect of gun ownership levels on violence rates.