Methodology

Updated: February 2021

Firearm Deaths

Fatal injury data comes from the Centers for Disease Control and Prevention (CDC)’s National Vital Statistics System (NVSS). The CDC compiles this data from death certificates, which are provided by vital registration systems from state and local jurisdictions.

EveryStat uses CDC data from the Wide-ranging ONline Data for Epidemiologic Research (WONDER).

Version: The most recent data from the CDC published on EveryStat is for 2019. Unless noted, and in order to account for normal fluctuations that occur between years, deaths are a five-year average of 2015 to 2019.

 

Intimate Partner Homicides

Statistics of intimate partner homicides are based on data from the Federal Bureau of Investigation (FBI)’s Uniform Crime Reporting Supplementary Homicide Reports (SHR), which compile data from law enforcement agencies. This is one of the few sources of homicide data that includes information about the relationship between the victim and offender. 

Version: The most recent data from SHR published on EveryStat is for 2019. Deaths are reported as a five-year total from 2015 to 2019.

 

Fatal Police Shootings

Data on fatal police shootings comes from the Mapping Police Violence database. While the CDC is considered to be the most comprehensive source for data on fatal firearm injuries, its data on shootings by law enforcement is likely greatly underreported. This is largely due to missing information on death certificates about cause of death, which may result in the misclassification of intent as assault rather than a police shooting.

Version: The most recent data from Mapping Police Violence published on EveryStat is from 2020. In order to account for normal fluctuations that occur between years, deaths are a five-year average of 2016 to 2020.

 

Transgender Homicide

Data on the homicide of transgender and gender non-conforming Americans is gathered by Everytown for Gun Safety through media reports. In this analysis, “transgender” and “gender non-conforming” identities include any victim who lived or identified as a gender different than that which was ascribed to them at birth.

Version: The most recent data from the Everytown Transgender Homicide Tracker published on EveryStat is updated in real time.

 

Firearm nonfatal Injuries

Data on nonfatal firearm injuries is based on the Agency for Healthcare Research and Quality’s (AHRQ) Healthcare Cost and Utilization Project (HCUP) surveys of hospital discharges.

Everytown has worked with Ted R. Miller and the Pacific Institute for Research and Evaluation (PIRE), using multiple datasets, to calculate nonfatal injuries:

  1. Restricted access National Emergency Department Sample (NEDS) of emergency department discharges.
  2. Restricted access National (Nationwide) Inpatient Sample (NIS) of inpatient hospital admissions.
  3. State Emergency Department Databases (SEDD), including both restricted access and publicly available data through hcup.net.
  4. State Inpatient Databases (SID), including both restricted access and publicly available data through hcup.net.

Note that because of rounding and analysis done to create state-level estimates, the sum of the state estimates will not equal the national total. For further analysis on nonfatal injuries, please see the report. For further information on the methodology behind the nonfatal firearm injuries count, please see the Methodological Note

Version: The most recent data on nonfatal injuries published on EveryStat is from 2017, the most recent state-level estimates available from researchers.

 

Economic Cost of Gun Violence

Data on the costs of gun violence are constructed from the calculation: Cost per nonfatal injury or death multiplied by the number of injuries and deaths

Costs per fatal and nonfatal injury are based on research by Ted R. Miller and PIRE. Costs to society include out-of-pocket payments for medical care, mental healthcare, emergency transport, police response, criminal justice and incarceration, employer costs, work loss for victims and perpetrators, and quality of life lost. Costs to taxpayers include government payments for medical care, mental healthcare, emergency transport, police response, criminal justice and incarceration, Medicaid, and Social Security Insurance payments. Further methodology of unit costs are available from the Methods Note.

The number of nonfatal firearm injuries used to calculate the total cost of gun violence for states was estimated from the 2017 Healthcare Cost and Utilization Project (HCUP) data on hospital emergency department and inpatient files (see Methodological Note). The number of firearm deaths used to calculate the total cost of gun violence for states was as per the CDC WONDER for 2018. 

Version: The most recent data on costs of gun violence published on EveryStat are in 2018 US$, based on 2018 firearm deaths and 2017 firearm nonfatal injuries.

 

Congressional District Analysis

Everytown calculated suicide by congressional district using county-level data on the number of suicides and firearm suicides obtained from CDC WONDER, Underlying Cause of Death combined with demographic data from the US Census Bureau. Allocation of counts from the county to the congressional district and calculation of rates for the congressional district was based upon the methods from the American Cancer Society and published on the CDC website for calculation of congressional district cancer mortality. Counts of suicides and firearm suicides for the congressional districts were calculated by weighting county counts by the proportion of the county population allocated to the congressional district for that population group. The weighted counts were then summed over the counties in the congressional district. For further information, please see the Methods Note.

Version: The most recent data on firearm suicides by congressional district published on EveryStat are five-year averages for 2014 to 2018 using the district boundaries for the 116th Congress (2019 to 2020).

 

Population Denominators

  • For national, state, and county-level deaths, population denominators are from CDC’s WONDER according to its bridged race and ethnicity categories.
  • For nonfatal injuries, population denominators are provided by HCUP.
  • For city and congressional district data, intimate partner homicides and fatal police shootings, population denominators are from either the Census (including mid-year estimates) or the American Community Survey, both accessible through https://data.census.gov/.

Suppressed and Missing Data

Updated: December 2020

In order to protect individual identities, the CDC and AHRQ do not report data when the number of deaths or nonfatal injuries is less than or equal to 10 for any subnational estimates. Further, the CDC considers population rates unreliable (large coefficients of variation) when the number of deaths or nonfatal injuries is less than or equal to 20. Data that is suppressed or considered unreliable has not been included in EveryStat.

The FBI’s SHR is voluntary reporting by local law enforcement agencies. As a result, reporting may be incomplete or missing for some states in some years and therefore could not be included in EveryStat.

Demographics

 

Updated: December 2020

Racial and ethnic categories in EveryStat are defined by the CDC. EveryStat contains five separate, mutually exclusive racial and ethnic categories.

American Indian = American Indian/Alaskan Native, non-Hispanic

Asian = Asian/Pacific Islander, non-Hispanic

Black = Black, non-Hispanic

Latino (sometimes self-identified as Latinx) = American Indian Hispanic + Asian Hispanic + Black Hispanic + white Hispanic

White = white, non-Hispanic

Children and teens are defined as individuals between the ages of 0 and 19 years. In statistics regarding the “leading causes of death,” however, children and teens are defined as those between the ages of 1 and 19. This is because the leading causes of death for newborns and infants less than 1 year of age are considerably different from those 1 year and up.

Analysis and Calculation

 

Updated: December 2020

Crude rate per 100,000 people = cases / population x 100,000 Crude rates per 100,000 people were calculated by dividing the number of cases by the population denominator and multiplying by 100,000.

Age-adjusted rates per 100,000 people Age-adjusted rates per 100,000 people were obtained directly from the CDC and not calculated by Everytown.

Disparity ratio = rate of population A (higher rate) / rate of population B (lower rate) To compare the rate at which a given outcome affects two different groups, a disparity ratio was used. For example, a disparity ratio was used to calculate the following: Black women are two times as likely as white women to be fatally shot by an intimate partner.

Percent change = (new value – previous value) / previous value Percent change is used to compare new values to previous values. This formula is used when looking at the percent increase or decrease.

Percent difference = (value A – value B) / average of values A & B Percent differences are used when comparing counts or rates across states (or other subnational geography) or intent.