Status of the database

Due to the significant resource requirements for maintaining this dataset, the Stanford Mass Shootings in America database has been permanently suspended. The project has been archived and will remain publicly available as a GitHub repository.

Alternatives to the MSA Database

As a service to those seeking a comparable and more recently updated alternative, we are providing links to other, more active efforts, below. The Stanford Geospatial Center in no way warrants the quality of, nor endorses any of the database efforts linked to in the following list and provides the links only as a starting point to what should be a measured and careful selection process when seeking data on phenomena of this type.

Mother Jones

https://www.motherjones.com/politics/2012/12/mass-shootings-mother-jones-full-data/

• Data Fields: City/state, date, brief description, fatalities, injured, venue, mental health history, weapons obtained legally, where obtained, type of weapon/details, race, gender, lat/long
• Data coverage: 1982-2017; updated frequently. Database contains 91 entries as of 10/12/17. 34% entries from 1982-1999, 34% entries from 2000-2012, 31.8% entries from 2013-2017
• Methodology:
  • Journalists are doing the research and curating the entries
  • Minimum sources cited for each incident: usually 3+ sources, with specific citation for mental health background of shooter, if applicable
• What’s included:
  • 1982-2012: a single incident, in a public place, resulting in 4+ victims killed by the attacker
  • 2013 onwards: a single incident, in a public place, resulting in 3+ victims killed by the attacker (adjusted in response to 2013 federal mandate pressing the investigation into incidents with this threshold)
• What’s not included: shootings stemming from “more conventional crimes” (armed robbery or gang violence), domestic violence

Gun Violence Archive

http://www.gunviolencearchive.org/mass-shooting

• Data Fields: Date, street address, city, state, # killed, # injured, sources
  • Individual incident webpages contain info not in the exportable CSV table: lat/long, People (victims and suspects) and their names, age, gender, and status (arrested, killed, etc), and district numbers for congressional, state, and house.
• Data Coverage: 2013 - Present (plus 3 records from 2012), updated daily. 1522 records as of 10/12/17. Even data spread.
• Methodology :
  • Dedicated, professional staff which sweeps over 2,500 law enforcement and media resources daily, logging incidents with up to 120 potential incident variables.
  • Utilize automated queries and manual research. Each incident is verified by both initial researchers and secondary validation processes.
  • Minimum sources cited for each incident: 1+
• What’s included:
  • Mass Shooting definition = 4+ shot and/or killed in a single incident , at the same general time and location, not including the shooter. October 2017
  • GVA separately catalogs comprehensive gun violence stats by category, including total number of gun injuries, victims of an armed robber or home invader, incidents of defensive gun use, store clerks who stop a robbery, individuals who stop an assault or rape with a gun.
• What’s not included: victim or subject-suspect characteristics such as race, religion, mental health status, nationality, or political affiliations.

Mass Shooting Tracker

https://www.massshootingtracker.org/data

• Data Fields : date, name of shooter, number killed/wounded, city/state, and sources
• Data coverage: Jan 2013 - Present. Updated daily. Database contains 1861 entries as of 10/12/17. Even data spread, usually one or more entries per day.
• Methodology:
  • Unfunded, crowd-sourced effort. Users can report a mass shooting by submitting sources to the moderators of the subreddit that manages the database. (Note: must have Reddit account to submit.)
  • Minimum sources cited for each incident: 2+
• What’s included:
  • an incident of violence in which 4 or more people are shot (not necessarily killed). Injured count may include the gunman or police shootings of civilians around the gunman.
  • Shooter death (on scene or soon after the fact) is included in fatality count.
• What’s not included: motive of the shooter, venue of the incident (school, workplace, street), mental health history of shooter.

MSA (Mass Shootings in America)

Status of the database

Due to the significant resource requirements for maintaining this dataset, the Stanford Mass Shootings in America database has been permanently suspended. The project has been archived and will remain publicly available as a GitHub repository.

A repository for the maintenance and distribution of the Mass Shootings in America Database.

Contents

Data Folder contains the most updated MSA dataset as .csv and GeoJSON

Methodology Folder contains the rules for how events are included into the dataset and how it is organized.

Project background

The Stanford Mass Shootings of America (MSA) data project was begun in 2012 in reaction to the mass shooting in Sandy Hook, CT. In our initial attempts to map this phenomena it was determined that no comprehensive collection of these incidents existed online. The Stanford Geospatial Center set out to create, as best we could, a single point repository for as many mass shooting events as could be collected via online media. The result was the Stanford MSA.

What the Stanford MSA is

The Stanford MSA is a data aggregation effort. It is a curated set of spatial and temporal data about mass shootings in America, taken from online media sources. It is an attempt to facilitate research on gun violence in the US by making raw data more accessible.

What the Stanford MSA is not

The Stanford MSA is not a comprehensive, longitudinal research project. The data collected in the MSA are not investigated past the assessment for inclusion in the database. The MSA is not an attempt to answer specific questions about gun violence or gun laws.

The Stanford Geospatial Center does not provide analysis or commentary on the contents of this database or any derivatives produced with it.

Data collection methodology

The information collected for the Stanford MSA is limited to online resources. An initial intensive investigation was completed looking back over existing online reports to fill in the historic record going back to 1966. Contemporary records come in as new events occur and are cross referenced against a number of online reporting sources. In general a minimum of three corroborating sources are required to add the full record into the MSA (as many as 6 or 7 sources may have been consulted in many cases). All sources for each event are listed in the database.

Due to the time involved in vetting the details of any new incident, there is often a 2 to 4 week lag between a mass shooting event and its inclusion in the public release database.

It is important to note the records in the Stanford MSA span a time from well before the advent of online media reporting, through its infancy, to the modern era of web based news and information resources. Researchers using this database need to be aware of the reporting bias these changes in technology present. A spike in incidents for recent years is likely due to increased online reporting and not necessarily indicative of the rate of mass shootings alone. Researchers should look at this database as a curated collection of quality checked data regarding mass shootings, and not an exhaustive research data set itself. Independent verification and analysis will be required to use this data in examining trends in mass shootings over time.

Definition of Mass Shooting

The definition of mass shooting used for the Stanford database is 3 or more shooting victims (not necessarily fatalities), not including the shooter. Identifiably gang, drug or organized crime related shootings are not included in the database.

Why limit the type of incidents added to the Stanford MSA?

The goal of the Stanford MSA is to track the particular phenomena of mass shootings in the U.S. and not gun violence as a whole. While all gun violence is tragic, it is the seemingly spontaneous shooting incidents that are the most confounding when looking for answers about motivation, after the fact.

Differences between the Stanford MSA and other mass shooting data projects

The key difference between the Stanford MSA and other similar projects is the scope of definition and the focus on mass shooting incidents versus mass murder. Instead of limiting our data collection to incidents in which 4 or more fatalities occurred (the FBI definition for Mass Murder), we instead collect incidents of 3 or more shooting victims (not necessarily fatalities). All mass shooting definitions are arbitrary in that there is no natural way to quantify such an event. In setting up our threshold we wanted to focus on the shootings over the outcomes.

How to compare the Stanford MSA totals to other project totals for 'Mass Shootings'

There are a number of other databases tracking shootings in the U.S. each with their own criteria and incident tally. Extreme differences between datasets are likely an indication of different thresholds for inclusion (1, 2, 3 or 4 victims), the type of outcome (fatalities only versus fatalities and injuries), and the type of incidents captured (all gun violence, gang related, domestic, terroristic, etc.).

Known issues

• Data for the Stanford MSA are collected and maintained with the help of student assistants, interns, or temporary staff. The database may be left untended for short periods between staff hires during which events may go unrecorded and remain unremediated when work on the MSA resumes.

• Locations in the MSA are generally given as City/State. When using some online mapping services to geocode locations by City/State some locations may be misplaced due to the repetition of common city names between multiple states. This is an issue with the mapping service, not the database. Double check any maps created with this data to be sure your chosen service is placing locations correctly.

• Sources for the MSA come solely from online reports. Increases in records over time can not simply be assumed to be a reflection of increased events as online reporting trends and outlets have increased tremendously over the span of time covered by the Stanford MSA.

• Some shooting incidents defy straightforward classification. Ambiguous records are flagged with a value of 2 in the depreciation field. For a more detailed explanation, see data dictionary in the Methodology folder.

A note to researchers and journalists using the MSA

It is the user's responsibility to perform their own rigorous data quality assessment before using the MSA in their work. Please read the data dictionary completely when looking for incident breakdown information. Please refer to the Known Issues and Methodology sections above for information on other issues that may affect trends in the data. Graphing the raw data for commentary without adjusting for the various known issues and methodology (see above sections) may not provide insightful information.

How to cite the MSA

The Stanford MSA is released under a Creative Commons Attribution 4.0 international license. Please cite the MSA as “Stanford Mass Shootings in America, courtesy of the Stanford Geospatial Center and Stanford Libraries”.

Status of the database

Due to the significant resource requirements for maintaining this dataset, the Stanford Mass Shootings in America database has been permanently suspended. The project has been archived and will remain publicly available as a GitHub repository.