Associations between firearm and suicide rates: a replication of Kleck (2021)

Objective In a recent study using data from 194 nations and 28 EU member states, Kleck (2021) reports that firearms are associated with the rate of suicides committed with firearms, but not non-firearm suicides or suicides overall. He concludes that firearm accessibility only influences suicide method. However, this conclusion is unjustified because the findings are internally inconsistent: if firearms only influence firearm suicide rates, they should be negatively associated with non-firearm suicides to reflect method substitution. I attempt to replicate Kleck’s results and, if there are differences, offer potential explanations. Method I reproduce Kleck’s analytical approach: ordinary least squares regression with log-transformed suicide rates as the dependent variable and crude firearm rates as the independent variable, weighted by the square root of the population. I also test ranked and z-scored firearm rates to account for outliers and check robustness of results, stratify analyses by various demographic subgroups, and include US state data. Results Firearms were positively associated with total suicide rates globally and within the US but not in the EU. Across all geographic units, firearms were associated with firearm suicides but not non-firearm suicides. Associations were robust to modifications of the firearm rate, though varied within subgroups. Firearms were negatively associated with non-firearm suicides among older persons and black Americans, indicating method substitution. Firearms were also negatively associated with total suicides among older persons in global analysis, which may be due to depletion of susceptibles at younger ages where firearms are more readily available. Conclusion Kleck’s findings failed to replicate. Therefore, I reject his conclusion that firearms do not influence suicide.