Abstract
The United States combine high rates of firearm homicides with high gun prevalence. In the past, a significant positive association was found between the two. This study revisits the gun prevalence-gun homicide debate using more elaborate estimates of gun ownership for the 50 States. Longitudinal data (1999-2016) were analysed with Bayesian multilevel Gamma-Poisson models. The results demonstrated a very small positive association that diminished after adjusting for crime rates. Findings suggest that the association either attenuated in more recent years, or previous studies had overestimated this association.