Citation

Clark DA, Macinko J, Porfiri M. Soc. Sci. Med. (1982) 2022; 305: e115103.

Copyright

(Copyright © 2022, Elsevier Publishing)

DOI

10.1016/j.socscimed.2022.115103

PMID

35696874

Abstract

Guns are a ubiquitous feature of contemporary US culture, driven, at least partly, by firearms' constitutional enshrinement. However, the majority of laws intended to restrict or expand firearm access and use are formulated and passed in the states, leading to 50 different firearm-related legal environments. To date, little is known about why some states pass more restrictive or permissive firearm laws than others. In this article, we identify patterns of firearm law adoption across states, by framing the problem as a bipartite network (states connected to laws and laws connected to states) that is the result of a complex, and interconnected system of unobserved forces. We employ Exponential-family Random Graph Models (ERGMs), a class of statistical network models that allow for the dispensing of the assumptions of statistical independence, to identify factors that increase or decrease the likelihood of states adopting permissive or restrictive firearms laws over the period 1979 to 2020.

RESULTS show that more progressive state governments are associated with a higher chance of enacting restrictive firearm laws, and a lower chance of enacting permissive ones. Conservative state governments are associated with the analogous reversed association. States are more likely to adopt laws if bordering states have also adopted that law. For both restrictive and permissive laws the presence of a law in a neighboring state increased the conditional likelihood of a state having that law, that is laws diffuse across state borders. High levels of homicides are associated with a state having adopted more permissive, but not more restrictive, firearm laws. In summary, these results point to a complex interplay of state internal and external factors that seem to drive different patterns of firearm law adoption Based on these results, future work using related classes of models that take into account the time evolution of the network structure may provide a means to predict the likelihood of future law adoption.

Language: en

Keywords

Violence; Social networks; Public health; Policy analysis; Government/state