Gunshots Project Data: Gun Deaths In Vermont, 2011-2016
VPR compiled a database of all of the gun deaths in the state of Vermont between the beginning of 2011 and the end of 2016. The database has information about every person who died by gunshot during that period: including age, veteran status, marital status, education level and other details that can help Vermonters better understand which parts of their community are most affected by gun deaths.
UPDATE May 22, 2018
In August 2017, this project originally reported the Vermont gun deaths from 2011-2016. In May 2018, VPR collaborated with the St. Albans Messenger to add the records from 2017 using the same methodology described below. The datasets are here on separate tabs. Read and listen to VPR’s “Gunshots Project Update: Takeaways From 2017 Gun Deaths Data.”
UPDATE June 4, 2019
The project has now been updated with gun deaths data from 2018; that dataset is available here on its own separate tab. Read and listen to VPR's "Gunshots Project Update: Takeaways From 2018 Gun Death Data."
ABOUT THE DATA
The data source: The data come from the Vermont Department of Health.
The data set: The data set for this project consists of all deaths caused by gunshot wounds in the state of Vermont between Jan. 1, 2011 and Dec. 31, 2016. (In a few cases, a person who sustained a gunshot wound in a different state later died at a Vermont hospital. Those cases are included in this data set.)
What the data show: All of the data on the Death Certificate Data Input originated from death certificates provided by the Vermont Department of Health. We have not added any new information or context to what was provided by VDH.
What the data don’t show: There are important factors relevant to a person’s life and death that don’t show up in the data we have. These include:
- Recent changes in employment or marital status (we know the status, but not when it changed)
- Toxicology at time of death (was the person under the influence?)
- Identity/motives of killer. In homicides, the death certificate doesn’t name killer(s) or motive(s)
- Criminal justice outcomes. We don’t know which murders are unsolved, which ones have led to criminal sentences, or what those sentences are.
- How overall trends relate to national averages and/or the general population of Vermont
- We know the average age of people who are killed by guns in Vermont, but that’s only meaningful when compared to the average age of all people in Vermont.
- We know what percentage of gun deaths are suicides and what percentage are homicides, but that’s more meaningful when compared to national averages. In cases like these, more reporting using outside data will be required.
Data quality: Here is how the we put the database together:
- Requested and received, from the VT Department of Health, paper death certificates for each of the gun deaths in a five-year period.
- Scanned all 421 pages and converted them into five-page batches (one six-page batch) and sent those PDF files to volunteers on staff and not on staff
- Volunteers transcribed data from the scanned death certificates into a Google form, and those responses went directly into our database.
- Confirmed that every page had been entered into the database
- VPR’s Taylor Dobbs and Sara Simon audited the data entry. With Taylor looking at the scanned PDF versions of death certificates, and Sara looking at the database, we read aloud certain data points from each death certificate (those are the bolded columns in the master spreadsheet) and confirmed that the database matched the documents.
The database may contain typos or incomplete words. If that is the case, it’s because the death certificate also contained those errors. We have only confirmed that the database matches the source documents - we have not independently reported out any of these details.
Data withheld: Before publicly releasing the data, VPR staff removed personally identifying details from the data set. While all of the information in the database comes from public records, journalistic ethics require VPR to consider the implications of publishing the full names and addresses of the people who died.
NPR’s ethics handbook states that “[e]veryone affected by our journalism deserves to be treated with decency and compassion.” The handbook goes on to say that “[w]e minimize undue harm and take special care with those who are vulnerable or suffering."
Out of respect for those who died, as well as their families and communities, VPR made the decision to de-identify the information in the database. Here are the data from the certificates we decided to withhold from our public database.
- Legal name - withheld.
- Aliases - withheld.
- Date of Birth - withheld.
- Residence: Town, State - Specific home addresses were withheld, Town & State remain.
- Spouse/Civil Union Partner's Birth Name - withheld.
- Father/Parent's Birth Name - withheld.
- Mother/Parent's Birth Name - withheld.
- Informant & Informant Address - withheld.
- Place of Death: person addresses were withheld, but descriptive text (e.g. “decedent’s home”) remains. Town and state remain in the public data set, as do addresses of public places (e.g. “Emergency room/Outpatient, Copley Hospital, Morrisville, VT”).
- Injury Location: Town and State remain, specific addresses were withheld.
Questions, comments, corrections: If you see an error in the data or have a question, comment or request, please contact the VPR News team.