Firearm Mortality and State Characteristics

Rates of homicide, death by suicide, and firearm violence vary dramatically among U.S. states. The interactive tools on this page allow users to investigate how these differences correspond with state characteristics and to answer questions, including “How much higher is Missouri's firearm homicide rate than what would be expected based on other states with similar economic conditions?,” “Are states' firearm ownership rates related to differences in states' suicide rates?,” “Are firearm suicide rates more strongly associated with state firearm ownership rates than with state population demographics?”

The tool allows exploration of the following types of state characteristics: population demographicsPopulation demographicsAge, sex, and race and ethnicity, economic conditionsEconomic conditionsIncome, poverty, unemployment, and education, geographic conditionsGeographic conditionsPopulation density, region of the country, urbanicity, and latitude, firearm prevalenceFirearm prevalenceHousehold firearm ownership rates and share of suicides and homicides committed with a firearm, and the political climatePolitical climateVoter turnout and political party composition of elected officials. Each characteristic model includes all available measures.

Select a mortality measureMortality measuresTotal firearm deaths, firearm suicides, firearm homicides, total suicides, and total homicides and a set of state characteristics from the drop-downs below. The tool will compare mortality rates for each state with other states that are similar on the chosen characteristics. If you select a state, you will get additional information about mortality rates in that state and how it compares with similar states.

Associations between state characteristics and mortality alone do not imply that changes to those state characteristics would cause changes in state mortality rates. The associations could reflect a causal relationship, or they could be explained by other causes that affect both mortality and the state characteristics.

Data and Methods

All mortality estimates are constructed from the Centers for Disease Control and Prevention's (CDC's) Wide-Ranging Online Data for Epidemiologic Research (WONDER) data (for more details, see the tool documentation). CDC WONDER is a website hosted by the CDC that releases public data from its National Vital Statistics Surveillance system (CDC, undated-b). CDC privacy protections prohibit the disclosure of mortality rates that are based on fewer than ten decedents. Measures of state characteristics come from a variety of sources and are outlined in the tool documentation.

For each state, mortality outcome, and set of state characteristics, we estimate an average mortality rate for other states that are similar to the target state on that set of characteristics. To get these state-specific comparisons, we first estimate machine learning models to predict each mortality outcome from that set of state characteristics using data from the other 49 states. The average mortality rate for similar states is estimated as the out-of-training-sample estimate from that machine learning model, given the state characteristics of the target state. The r² estimates presented should be interpreted as cross-validated r².

For more details on the methodology and data sources, see the tool documentation.

References

• CDC—See Centers for Disease Control and Prevention.
• Midgette, Gregory, Terry L. Schell, and Andrew R. Morral, Visualizing Correlates of State Firearm Mortality, RAND Corporation, TL-A4360-2, 2026. As at January 29, 2026: https://www.rand.org/pubs/tools/TLA4360-1.html
• Centers for Disease Control and Prevention, “Underlying Cause of Death, 1999–2020,” WONDER data system, undated-b. As of August 17, 2022: https://wonder.cdc.gov/ucdicd10.html

Digital Credits

Kekeli Sumah (design), Nelson Correia and Lee Floyd (development), and Heather McCracken (project management)