Should we be Arming Teachers?
In America, school shootings never quite disappear from the news.
Yesterday, for instance, former deputy Scot Peterson went on trial for taking cover outside a Parkland, Florida high school building in 2018 while an active shooter was inside.
Yesterday also marked the 10-year anniversary of the mass shooting that took place on and near the Santa Monica College campus. Who remembers that, given the seemingly endless succession of tragedies since then? Yesterday, while still contemplating the first anniversary of the Uvalde massacre, we also learned that seven people were shot outside a high school graduation ceremony in Richmond, Virginia.
I could go on, but there's no need. You know the awful story. Gun violence in schools and other settings has become an indelible part of the American experience. No history of our country will be complete without it.
Arming teachers
Preventing school shootings, or at least responding to them more effectively, also gets a lot of attention. My focus will be on the practice of arming teachers. In 2013, following the Sandy Hook massacre, South Dakota became the first state to allow certain school employees to carry guns, an option that has since become (or is on the verge of becoming) legal in 34 states.
The purpose of this newsletter is to ask whether we should be arming teachers in the first place. Statistics play an important role in describing the scope of the problem, determining whether Americans support this strategy, and deciding whether it's effective.
The scope of the problem
By almost any definition of the term, America leads the world in school shootings.
The federal government doesn't maintain official statistics. In 1996, the so-called Dickey Amendment, named for Republican congressman Jay Dickey, prohibited the CDC from spending money that might "advocate or promote gun control", and for this reason we have no federal databases of direct relevance. What we know about school shootings comes from the work of independent researchers, journalists, and non-profit corporations. The most comprehensive databases are maintained by news organizations (e.g., The Washington Post) as well as a mix of partisan (e.g, Everytown) and non-partisan (e.g, Gun Violence Archive) non-profits.
For instance, from 2009 through 2018, U.S. educational institutions, including colleges and universities, experienced anywhere from 288 to more than 600 school shootings. Higher estimates reflect more inclusive definitions of school shootings. The largest estimates include situations where a gun is fired but no injuries are reported, or the shooting takes place after hours, etc.
Consider for a moment that widely-reported 288 statistic, which represents the number of incidents in which at least one person was struck by a bullet on campus. Here's some important context:
1. It's an underestimate.
Minor injuries, as well as victimless shootings, are among the incidents that aren't always recorded. Experts almost universally agree that the available figures underestimate the frequency of shootings.
2. It's a much higher number than for other countries.
Wherever reliable data are available, differences betwen the U.S and other countries are staggering. 288 shootings in U.S. schools between 2009 and 2018? There were less than 10 forĀ each of the other 27 countries that were studied during the same time period. This isn't just because we're a large country. Our per capita rates of school shootings are much higher as well.
Clearly, we have a serious problem, no matter how school shootings are defined.
Snopes.com, one of the most thorough and most credible nonpartisan fact-checking organizations, looked at the 2009-2018 data for other countries and identified some inaccuracies, but the extent of error was painfully small. According to Snopes, Canada had four shootings rather than two; Mexico had nine rather than eight.
Some reassurance
This is a fairly grim newsletter, so let's pause for a moment and take a breath. If you're a parent, an educator, or just a concerned citizen, please remember that the problem America has with school shootings doesn't translate into very imminent personal risk.
To illustrate the point, consider 2022, a record-breaking year for school shootings in our country. There were 46 shootings on K-12 campuses in which at least one person was injured or killed. Compared to other countries, that's a huge number. But there are more than 130,000 K-12 schools in the U.S, which means that less than 4 out of 10,000 experienced a shooting.
Causes
Sadly, the U.S. is the only country in the world in which there are enough school shootings to permit statistical analysis of possible causes.
At the same time, school shootings are still rare enough that teasing out the possible contributors is challenging. (See here for a technical discussion.)
The goal of statistical analysis in this case would be to identify predictors. We want to know where and when school shootings are most likely. We want to know when extra precautions or immediate intervention is needed.
People do a lot of profiling, both formally and informally, but a focus on individual shooters doesn't buy us much. Active school shooters, for instance, tend to be white males with a strong interest in weapons and violence. They "tend" to have a lot of other things too, including a history of emotional disturbances, peer rejection, and feelings of alienation. But knowing what they "tend" to be like doesn't help identify them in advance. The vast majority of people do not become active shooters. This includes white males who are emotionally disturbed, rejected, and alienated. We don't know yet how to pick out the fractionally tiny percentage of these individuals who open fire in schools (unless they've just communicated their plans to someone).
Stepping back a bit, the strongest predictors of school shootings tend to be those that predict violent crime in any setting. For instance, an influential 2019 study found that K-12 school shootings are most common in schools that have higher enrollments, greater levels of economic disadvantage, and higher rates of violent crime. In other words, school shootings can be viewed as one kind of violent crime that's more prevalent in certain kinds of schools. This tells us where extra support is needed, but it doesn't allow us to predict where incidents will occur. After all, in 2022, we could not have guessed which 46 out of the tens of thousands of K-12 schools that fit the description above would experience a shooting.
Finally, in spite of what you might hear from both sides of the gun-control debate, state-level analyses show that the strongest predictor of school shootings is simply the size of a state's population rather than its political leanings. California, a blue state, has the largest population and the most school shootings. Texas, a red state, has the second-largest population and the second-most shootings, and so on. (In Appendix A, I describe some further analyses I ran on state-level data and possible ideological or regional contributors.)
Failure to arm teachers is not a cause
On April 10, Congressman Thomas Massie (R-Kentucky) co-authored a short opinion piece in the Washington Times entitled "There have been no shooting attacks in schools where teachers can legally carry guns."
Rep. Massie and co-author provided no evidence to support their claim ā probably because there is no evidence. 34 states have passed (or are in the midst of passing) legislation that permits teachers and other staff to carry weapons to school under certain conditions. Since enacting said legislation, every one of those states has experienced at least one school shooting, according to any accepted definition of the term.
Rep. Massie was wrong. But we might ask whether a revised version of his claim makes sense. Are there fewer shootings in schools that permit armed teachers on campus?
It's impossible to tell at this point. Laws allowing teachers to bring firearms to school are relatively new in most states, and there haven't been enough shooting incidents to make analyses viable. An additional problem is that we don't have consistent data on how many (if any) teachers carry guns in schools where it's allowed.
In sum, Rep. Massie's original claim is wrong, and we can't say yet what influence, if any, armed teachers have on the incidence of school shootings. Time will tell.
What does the public want?
Another approach to deciding whether to arm teachers is to say: We live in a democracy. Let's find out what people want, and then respect their wishes.
Statistics are helpful in this regard. They tell us that, by a small margin, most people are opposed to arming teachers. This has been demonstrated again and again in large, nationally representative surveys, regardless of whether the respondents are adults in general, parents with children in public schools, or teachers themselves.
For instance, in a PDK International poll administered last year, roughly two weeks after the Uvalde shooting, 1,008 respondents were asked questions such as āWould you or do you support or opposeā¦allowing teachers or other school staff to carry guns in school?ā Response options included strongly support, somewhat support, neither, somewhat oppose, and strongly oppose. The table below sums up a few of the key findings:
Each percentage in this table tells you how many people either "somewhat" or "strongly" supported armed teachers or armed police in public schools. Here's some additional context:
1. The percentages in the table align closely with the results of multiple polls going at least as far back as 2018, though there's been a slight uptick in the percentage of people who support arming teachers.
2. The percentages are similar regardless of whether the respondents are adults in general or parents with children in public schools (e.g., 43% of the latter support arming teachers, compared to 45% of all adults).
3. People are slightly less likely to "strongly" support arming teachers than to "somewhat" support the practice. For instance, that 45% of adults who support arming teachers consists of 21% who are strongly supportive, and 24% who are somewhat supportive.
4. Arming teachers is supported by three times as many Republicans as Democrats. Other variables matter too (e.g., people who are less educated and/or live in the South are more supportive), but divisions are sharpest when considering party affiliation or ideology (liberal vs. moderate vs. conservative).
5. The majority of people, regardless of party affiliation, support the presence of an armed police officer on campus. Support is strongest among Republicans.
Overall, the majority of people oppose arming teachers, but it's a slim majority. Only a majority of Democrats and Independents feel that way. Nearly three-quarters of Republicans would like to see teachers bring firearms to school.
What do teachers want?
Two weeks ago, the Rand Corporation, a prominent, nonpartisan organization, released the results of a survey of 973 randomly chosen K-12 teachers. Teachers were asked, among other things, how allowing teachers to carry firearms for school security would affect school safety. Here's how they responded:
Only 20% of teachers felt that allowing teachers to carry firearms would make schools safer. This suggests even less positivity than the general public toward the practice. Even if every teacher who chose "Neither" supported the practice, that would still leave 54% of teachers who feel that armed teachers make schools less safe.
Unfortunately, Rand didn't ask teachers about political affiliation or ideology. The biggest subgroup difference was seen for school location:
We can't know whether or not the teachers who chose "Neither" would support the practice of arming each other, but in any case, it's clear that even among rural teachers, there's not strong support.
The Rand report contains much more data, but you get the idea. Among the general public, a slim majority is opposed to arming teachers. Among teachers themselves, opposition to being armed seems to be even stronger.
Is there direct evidence?
In theory, arming teachers could reduce school shootings in two ways: Through intervention (an armed teacher confronts a shooter and perhaps opens fire), or through deterrence (a shooter decides not to approach a school because he knows some of the teachers are armed).
There's no direct evidence yet as to whether arming teachers helps in either sense. This is partly because it's a relatively new practice, and partly because the incidence of school shootings is low, statistically speaking. Complicating matters are differences from state to state, district to district, and school to school in the nature of the policies, which continue to evolve over time. (Arkansas law provides a nice illustration of the complexities. Public K-12 employees are not permitted to have guns in school buildings, but they can keep them in their vehicles. There may be at least one gun on campus anyway, because, since 2021, operable firearms can be used by instructors teaching a hunting safety class to 5th through 12th graders. Meanwhile, churches that operate private K-12 schools can allow employees to keep guns at school. Whew.)
Here some things we do know:
1. Arming teachers costs time and money. Depending on state and local laws, schools may need to purchase gun safes, additional mental health screenings, regular trainings, liability insurance, etc.
2. Arming teachers has already resulted in over a hundred gun-related accidents in schools. We also know that when people carry guns, they sometimes shoot innocent people.
3. Arming teachers would not completely solve the problem. A Rand study showed that when gunfire is exchanged, even police officers only hit their targets about 18% of the time. Armed teachers, unless they have a military or law enforcement background, will be less well-trained than the police and thus unlikely to handle a shooter more effectively. Police generally receive around 700 hours of training, with a substantial amount of time devoted to firearms, along with 40 additional hours if they become resource officers and work at a school. Teachers receive much less training. In Ohio, for instance,Ā 24 hours of training is required. In Texas, the amount of training is left up to the school district.
Of course, you could argue that these costs are outweighed by the benefits of arming teachers. It might even be a persuasive argument. But it would an argument based on ethical and personal preferences rather than evidence, because we don't know yet how the incidence of school shootings is affected by the presence of armed teachers. (This newsletter focuses mainly on incidence. A separate question ā also unknown ā is whether armed teachers might reduce the severity of the shootings.)
In time, an accumulation of case studies may tell us how effectively armed teachers respond to active shooters. If school shootings decrease as armed teachers increase, we might assume a deterrence effect, but I'm skeptical that what we'd actually be observing is deterrence. This is a matter of logic rather than statistics. If someone is so profoundly disturbed that they would shoot up a school, but they choose not to do so because they know the teachers are armed, what happens next? They're still the same, profoundly disturbed person. Do they wait for a school bus to pass and shoot at it? Do they go to a restaurant and fire at patrons? This isn't deterrence. It's deflection. The armed teacher would only be sending a shooter somewhere else.
Is there indirect evidence?
Yes, in the sense that increasing civilian ownership of guns doesn't seem to prevent violence. Historically, the passage of right-to-carry laws is associated with a roughly 13 to 15 percent increase in violent crime over a 10-year period. As noted in Appendix B, the presence of armed school resource officers on campus does not result in fewer school shootings.
Although anecdotal evidence shows that a person with a gun can save lives by threatening or shooting an active shooter, on the whole the data suggest ā indirectly ā that arming teachers wouldn't reduce the incidence of shootings. (For details on the statistical challenges of linking gun control policies to outcomes, see here.)
Conclusion
Seven years ago today, four individuals were shot at Jeremiah Burke High School in Boston, less than five miles from where I'm sitting.
Nobody, other than long-time residents, remembers that incident. Locals are talking more about another shooting that took place at Burke last October.
The earlier shooting illustrates why arming teachers could only be, at most, a partial solution to a society-wide problem.
On June 8, 2016, while students were standing outside Burke for a fire drill, shots were fired into the crowd, killing 17-year-old Raekwon Brown, and injuring two other students as well as a 67-year-old pedestrian.
As it turned out, three gang members had been firing at a member of a rival gang, but missed. The injured pedestrian reported later that Raekwon Brown, who had no gang affiliations, saved her life by pushing her down once the shooting began.
Armed teachers could not have prevented this incident. At most, they or a school resource officer could've prevented further carnage ā if the gang members had decided to continue shooting. If armed teachers had been a deterrent, the gang members would've simply attacked their target elsewhere.
Once upon a time, in the late 20th century, we thought of schools as little islands in a sea of guns. We knew that America had a gun problem āĀ school shootings did occur sometimes ā but it wasn't until the 1999 Columbine massacre that most of us (apart from those already familiar with the most crime-ridden districts) started to recognize that schools are also vulnerable to the encroaching violence.
After Columbine came metal detectors, locked doors, security cameras, resource officers, etc. But, as we discovered from more recent incidents, metal detectors fail sometimes, locked doors may be left unlocked, security cameras break, and resource officers aren't always able, or willing, to help. Meanwhile the rate of school shootings has increased, and now we're arming teachers. Somehow, the hope persists that schools can become islands again amidst a sea of guns.
The problem is, schools aren't islands. They're just as much in the "sea" as any other structures. Students are shot in classrooms ā and on bikes and buses, in parking lots, at extracurricular events, on sidewalks during fire drills, at graduation ceremonies, and so on.
Again, the strongest predictors of school shootings are those that predict violent crime in any setting. Schools that are overcrowded, under-resourced, and serving high proportions of economically disadvantaged students need more counselors on campus, more access to social workers, more anger management programs, and other forms of support. Also needed are strategies that aren't school-specific. In a word, better gun control policies that apply in all settings will inevitably promote safer schools.
Thanks for reading ā and stay safe!
Appendix A: State-level analyses of school shootings
Using publicly-available data from 1970 through 2022, I ran some analyses to look more closely at state-level contributors to school shootings. In this database, a school shooting was defined as any incident on a K-12 campus where at least one person was either injured or killed by a bullet (suicides excepted).
My first analysis looked at the correlation between a state's ranking in number of school shootings reported between 1970 and 2022 and the state's ranking in population as of 2022. (If you're a stats person: I did a Spearman rank correlation with adjustment for tied ranks.) The correlation between number of shootings and size was 0.89. This is a very strong association. The larger the state's population, the greater the number of school shootings.
My second analysis considered the exceptions. I focused, arbitrarily, on states where the discrepancy between number of shootings and state population was at least 10 ranks in magnitude. For instance, Louisiana is the 25th largest state by population, but it has the 12th highest incidence of school shootings, so you could call that a 13-rank discrepancy. (This is cruder but quicker than running per capita analyses, which yield a comparable pattern of results.)
What I found was that for the most part, states with disproportionately high numbers of shootings are red states, while states with disproportionately low numbers of shootings are blue states.
Specifically, Alabama, Arkansas, Louisiana, Mississippi, and Maryland all had much higher numbers of shootings given their size ā these are all red states, except for Maryland. On the other hand, Massachusetts, Minnesota, New Jersey, and Arizona all had much lower numbers of shootings given their size ā all blue states, except for Arizona.
Can we conclude that the dominant political ideology of a state (assuming this is what voter behavior tells us) tends to influence school shootings?
Maybe, but not solely on the basis of these data. After all, the majority of red and blue states don't fit this pattern. For them, the key variable is still size: the larger the state, the greater the number of shootings. Simply describing a state as red or blue is a bit simplistic anyway, especially for states that have shifted one way or the other since 1970. And, something else about the exceptional states may have caused them to be exceptions. For instance, studies have documented higher per capita rates of gun-related deaths in red states. This in turn may or may not arise from the higher rates found in the the small towns that are more prevalent in red states. In short, school shootings may be greater wherever gun violence is greater, and this in turn may or may not reflect ideological and/or regional differences.
Finally, with respect to the exceptional states I mentioned, guns are prohibited in schools in Alabama, Louisiana, Mississippi, and Maryland, but not in New Jersey, Minnesota, Arizona, and Arkansas (sort of), so there's no evidence here that arming teachers either helps or hurt.
Statistically speaking, the small number of shootings makes any conclusions tentative at best.
Appendix B: What about school resource officers?
School resource officers (SROs) are typically sworn law enforcement officers who serve a policing function in a school. Their authority and responsibities differ from those of security guards. About half of U.S. K-12 public schools have an SRO on campus at least one day a week.
A study currently in press at the Journal of Policy Analysis and Management shows that the presence of SROs in K-12 schools has a number of positive impacts (e.g, reduced violence on campus) as well as negative ones (harsher punishments that disproportionately affect students who are Black, male, and/or living with a disability).
However, the study finds no impact of SRO presence on school shootings, in spite of access to data on 94,918 schools. A 2019 study also showed no association between the presence of an SRO and the severity of shootings.Ā
We can't generalize from SROs to armed teachers, of course. SROs tend to be better trained at handling situations involving guns, and, at any given moment during the day, they're freer to attend to safety issues than teachers, who must also focus on instruction. On the other hand, each school typically has just one SRO, but many teachers could be armed. All the same, the results of this new study hint at least that the mere presence of armed adults on campus may not reduce the incidence of school shootings. Guns in schools, whether in the hands of teachers or SROs, doesn't look yet like a helpful solution.