Sean Phillippi: Data links gun deaths to gun ownership, not mental health, video games

I have lived more than 90 percent of my life in two communities, and they are linked in a way that is both sad and tragic. Those places are Broward County and Dayton, Ohio.

The link between those regions, and too many others, is that they have been sites of recent mass shootings.

After the massacre at Marjory Stoneman Douglas High School in February 2018 in which 17 people were killed, I was connected to the mother of one of the murder victims because she wanted to honor her daughter’s memory by running for office.

As her campaign manager, I heard Lori Alhadeff speak several times about losing her daughter on that Valentine’s Day.

It was gutting and heartbreaking every single time she told the story, and I can’t imagine the pain of actually going through something like that.

Even though I don’t live in Dayton anymore, I still have family there and visit the area several times a year. My mother lives only about a mile from the Dayton neighborhood where nine people were shot dead Aug. 4.

The epidemic of gun violence in the United States has become very personal to me, and as a data scientist, I felt compelled to research the topic to see what I could figure out.

It turns out that there is no one fixed definition of what a mass shooting is — which can be problematic when analyzing data. Because the term “mass shooting” can be a moving target, the one consistent measure I used in my research was total gun deaths.

I looked at total gun deaths in developed nations to see if there was some factor or variable that could help explain why the United States has over 75 percent of the total gun deaths in the developed world when we account for roughly 30 percent of the population among those combined countries. When perusing the civilian gun ownership data, what stuck out most to me was that there are almost 400 million guns in civilian hands in the United States, which is more guns than people (393 million civilian guns versus 325 million people).

The United States’ 33 percent share, among developed nations, of total military and law enforcement gun holdings closely aligns with our total population. The United States’ share of civilian guns in the developed world is more than 80 percent, which is significantly higher than what it should be based on the total population of the country.

One of the most shocking things I found was that in the United States there are more than 70 guns in civilian hands for every one gun in the hands of law enforcement and military.

Because civilian gun ownership in the United States is as much of an outlier as gun deaths, seeing whether those two things had any connections seemed like a good starting point.

The results were that for every 10,000 guns in civilian hands in the developed world 9.3 to 9.6 people die annually because of a gun.

(Model is named Reg1 in the online workbook. This basic analysis showed that civilian gun ownership in the developed world accounted for 99.8 percent of the variability in the number of gun deaths. That number was 75.3 percent when running the model without the U.S. — Reg3 in the online workbook.)

There is more than a 99.9999 percent chance that the relationship between civilian gun ownership and gun deaths in developed countries is positive. The total population was not a significant factor in these analyses when taking a comprehensive look at gun deaths in developed countries.

I then tested video games using total video game revenue by country and mental health using the rate of mental illness by country. Neither factor was significant in trying to predict or model gun deaths in developed countries. The only significant factors out of everything tested were civilian held guns and guns held by law enforcement, and when digging deeper over 95 percent of gun deaths in the United States are associated with civilian held guns.

(I calculated this by taking the coefficients of the number of gun deaths associated with law enforcement guns — 9.448 per thousand guns — and civilian guns — 9.219 per 10,000 guns — and multiplying it by the number of guns in those categories in the U.S. The math works out to 960 expected gun deaths for law enforcement guns and 36,263 expected gun deaths for civilian guns.)

I share a sentiment with the character of Sam Seaborn from The West Wing. In an episode of the show that first aired in 2000, Seaborn said, “I am so off-the-charts tired of the gun lobby tossing around words like ‘personal freedom’ and no one calling them on it. It’s not about personal freedom, and it certainly has nothing to do with public safety. It’s just that some people like guns.” To be clear, there is nothing wrong with liking guns. I like guns and enjoy shooting them at gun ranges. Despite this, the question I ask myself is, “Would I sacrifice a family member or loved one for a gun?”

The unequivocal answer is no.

I posit that question because if gun deaths increase with the number of guns in civilian hands, as illustrated above, then contributing to the increase in civilian guns could conceivably contribute to gun deaths.

I then, morally, have no right to ask other families to sacrifice people they love so I can have a gun when I am unwilling to make that sacrifice.

The threat of gun violence is ever-present in poorer communities and communities of color. If you think you can escape the gun carnage, then you should know one reason the Alhadeff family chose Parkland when moving to Florida was because, like everyone else, they viewed it as an extremely safe community, and it was … until it wasn’t.

Researchers have looked for ways to reduce the number of gun deaths. A Boston University article notes, “research … suggests that three laws implemented in some states could reduce gun deaths by more than 80 percent if they were implemented nationwide. Laws requiring gun identification through ballistic imprinting or microstamping were found to reduce the projected mortality risk by 84 percent; ammunition background checks by 82 percent; and universal background checks for all gun purchases by 61 percent.”

Statistician George Box once said, “All models are wrong, but some are useful.”

My models aren’t perfect, but I hope they are useful. The one recurring thought I had when doing this research is that there isn’t nearly enough research or data regarding this issue.

A study of gun violence research funding found that gun violence had less than 5 percent of the funding and volume of publications than it should be based on the number of people who die from gun violence.

Another old saying goes, “the first step in solving any problem is recognizing there is one.”

Guns kill people, and the data conclusively shows that, among developed countries, gun deaths go up as civilian gun ownership goes up.

Gun violence is a serious problem that literally has life or death consequences.

Every year, until we take serious and concerted steps toward doing more robust research and letting the evidence dictate strategies to reduce the number of gun deaths in this country, we will continue to pay for our inaction with tens of thousands of people’s lives.

___

Sean Phillippi is a data scientist and Democratic consultant based in Broward County. He has consulted on campaigns at the federal, state and local levels. He has worked for multiple Super PACs, the Florida Democratic Party, a Democratic gubernatorial campaign and several members of Congress. He is the Managing Member of TLE Analytics LLC, the political data and consulting firm he founded in 2012, as well as a graduate student pursuing his M.S. in Analytics at Georgia Tech.