Ever since the school shooting in Newtown, one of the preferred tactics for gun control advocates has been to argue that guns kill lots of children every year. It’s one hell of a propaganda claim, since there’s nothing quite like the bodies of dead children to empower the Civilian Disarmament Industrial Complex, override the analytical abilities of the voting public and try to sneak through a little gun control legislation. We’ve seen this tactic before from the New York Times and we thoroughly debunked it. Now it looks like a medical student and his advisor have taken it upon themselves to pick up the mantle and try to advance the party line a little further. And once again, they’ve used some remarkably flawed data to back up their efforts . . .
The study was conducted by a Boston medical student and his adviser at Harvard. They ostensibly investigated the death rate of “children” at hospitals from gunshot wounds. Right away we run into a problem, because like the New York Times article the medical student in question includes all patients under the age of 20 in his definition of “children.”
The reason the word “children” is so loaded is that it evokes images of innocent, cherub-faced little tots barely old enough for elementary school. The picture that the title of the article and that the study tries to paint is one of innocent little Suzie being shot and dying before her 10th birthday. But in reality, as I discussed in the Times takedown, the vast and overwhelming majority of deaths from firearms for this age range happens in those 15 and over.
Personally, my cut-off for calling someone a child is where the state believes that they’re capable of operating a deadly machine at high speeds on the public roadways: 16. Any reasonable person might expand that definition to, at most, 18. But the study included, ahem, “children” up to the age of 20 because the probability of being wounded or killed by a gun increases drastically for every year between the ages of 15 and 20. The numbers simply weren’t big enough using actual children, so the study’s authors padded their numbers with those cohorts most associated with gang membership.
OK, so the stats are plainly padded. But before we actually get to the numbers, where do these authors get their statistics? From the NBC article:
Madenci, and his colleague, Dr. Christopher Weldon, a surgeon at Boston Children’s Hospital, tallied the new statistics by culling a national database of 36 million pediatric hospitalizations from 1997 to 2009, the most recent year for which figures are available.
Wait. So, what they were looking at were the numbers of pediatric patients hospitalized for gunshot wounds? That would exclude anyone who A) was admitted to a non-pediatric service, B) was killed at the scene and never transported to the hospital, or C) was treated at a hospital that didn’t report in to the database. That’s like trying to get water out of a well using a sieve.
OK, so the data set includes an abnormally large swath of the population and uses a terribly flawed database for its input. What were the results?
During that period, hospitalizations of kids and teens aged 20 and younger from gunshot wounds jumped from 4,270 to 7,730. Firearm deaths of children logged by hospitals rose from 317 in 1997 to 503 in 2009, records showed.
That doesn’t make any sense whatsoever. The authors claim that nationwide, the trend for “children” dying from firearm related causes is on the rise, but I just ran the numbers from the CDC, and they paint a different story.
For actual “children” (12 and under), the following stats were recorded:
- 1997: 318 fatalities
- 2009: 209 fatalities
For “children” between 13 and 19, here’s what the CDC said was going on:
- 1997: 3,905 fatalities
- 2009: 2,502 fatalities
Two things pop out immediately. First, the study conducted by the med student and his mentor missed a couple thousand fatalities. Second, their conclusions were completely bogus.
If the study and NBC News are to be believed, the number of children dying from gunshot wounds is on the rise. But if we look at an actually unbiased source (the CDC, who collect all death-related statistics in the United States), the number of deaths from firearms is declining.
What’s even more astounding is that the raw number going down, not just the rate. With the increase in population we might expect that the raw number of children killed by guns to rise in proportion to the population, but the rate to remain the same or drop slightly. In this case, both the raw number AND the rate are declining. And yet NBC News and Discovery are both reporting that these numbers are on the rise.
The issue here is really the source of the data. If the authors wanted to know the truth about whether increased gun ownership is increasing the risk of fatal incidents involving firearms, the CDC’s numbers are there and readily available — he didn’t need to go trudging through an incomplete and obscure database for statistics. But rather than look at the whole picture, they decided to restrict their input to deaths in hospitals.
As an EMT, I can tell you that we don’t typically transport corpses to the hospital emergency room — they go straight to the morgue. Those instances aren’t counted in the study. Similarly, not all injuries are serious enough for a $400+ ambulance ride and $1,000 hospital bill. Those instances aren’t in the study either. And including both those instances, you start to clearly see the downward – not upward – trend in “children” and firearms related deaths.
In the Discovery version of the article, the author of the study officially loses all credibility as far as statistical analytical ability is concerned.
“Based on our research, we know that there is a clear correlation between household gun ownership (and gun safety practices) and childhood gunshot wounds in the home on a large scale,” Madenci said in an email to Discovery News. [...] He said he decided to look at the question of gun ownership and childhood gun deaths after the shooting at Sandy Hook Elementary School in 2012.
If the author were to say that the overall accident rate for children in houses with guns had increased, that would be an interesting finding. But he didn’t — he wrote that gun accidents happen more when guns are present. That’s like saying households with cars are more likely to have a family member die in a car accident. What would be a persuasive argument to me is if the author proved that having a gun in the home increases the overall death rate (including all methods of death), because that would indicate that guns increase the overall probability of death. But he didn’t.
All he did was point out that by owning a certain object you’re more likely to be killed by that object that means nothing. Unless he’s trying to get some free publicity to help in his search for an intern position at a hospital in the near future, that is.