One of the more popular reactions from the right to the Parkland, Fla, school shooting — and from Fox News personalities in particular — is that we need to put more cops in schools. Or to put in armed guards, or retired soldiers, who would in all likelihood be given the same powers as police.

Would it work? Probably not. But we first need to acknowledge something before delving into this discussion: a classroom is just about the safest place a kid can be in America. I realize that’s a difficult thing to accept after an event as traumatic as Parkland. According to a New York Times review of statistics from the Gun Violence Archive, 43 people were killed on school campuses in 2017, 25 were killed in 2016, and 33 in 2015. That includes both victims of mass shootings and conventional homicides (my somewhat-awkward term for homicides other than those from mass shootings). That’s out of about 55 million students attending K-12 schools, public and private. The odds of a given child getting killed in a mass school shooting — or any school shooting — are literally less than 1 in 1 million. The criminologist James Alan Fox points out that since 1990, there have been 22 shootings at schools in which two or more people were shot, or well less than one incident per year. That’s less than one incident per year out of 100,000 public schools and 33,000 private schools. This means that the average elementary, middle, or high school can expect to see a mass shooting about once every 150,000 years.

These shootings are of course incredibly traumatic. That trauma is amplified by the round-the-clock coverage they receive. And, of course, I’m sure that if you know someone who has been killed in a school shooting, these statistics probably mean very little to you. But good public policy should be driven by sound data, not collective trauma. Putting retired soldiers or TSA-like checkpoints in our schools to prevent mass shootings is a clumsy and incredibly heavy-handed solution in search of a problem.

But let’s get back to the important question: Does it work? To answer that question, we need to look at the costs and benefits.

On the benefits side, the answer is pretty unsatisfying — it’s just hard to say if there are any. Part of the reason for this is that, again, violent crime in school is already incredibly low — so low, that looking at actual violent crime statistics is basically useless. School shootings are so rare, there’s just no way to gather a sample size large enough to draw any conclusions. Anecdotally, Marjory Stoneman Douglas High School did have a sheriff’s deputy on duty at time of the shooting, but he never discharged his weapon. Would it have been different if there had been four or five or six armed guards? Possibly. But again, we simply don’t have enough of these incidents (thankfully) to say for sure. There’s also a question of public resources, here. Does it make sense to pay a half dozen armed guards to patrol your local school if that school isn’t likely to see a mass shooting for tens of thousands of years?

If we look at nonviolent crime, we do start to see enough incidents to possibly make some comparisons, but because of the nature of the questions we’re asking here, this data can also be pretty confounding. Studies have shown that schools with school resource officers (the common euphemism for cops who are assigned to schools) tend to have more nonviolent crime, but that may simply be because schools that see a lot of crime are the schools local officials believe need SROs. It also seems likely that those schools may not actually have more nonviolent crime, it’s that with SROs around, more of it gets reported.

Because of these problems with the data itself, some researchers instead survey how safe students feel in schools with and without SROs. But this method also comes with problems. Whether or not students feel safe could be the product of any number of things that bear little relation to how safe they actually are. I would imagine, for example, that students across the country feel less safe at school today than they did last week, even though their actual relative safety probably hasn’t changed. This of course is entirely understandable. But we want to make policy that creates actual security, not security theater. Surveys have also shown that black students tend to feel less safe in schools with SROs, while white students tend to feel more safe. This could simply be because black students are more distrustful of law enforcement. It could also be because black students are more likely to live in schools in areas with higher crime rates. It tells us little about the efficacy of SROs.

So the benefits are unclear, and possibly unknowable. But at most they are probably minimal, given how little violent crime takes place at school in the first place. What about the costs?

It seems safe to say that putting officers in schools definitely comes with some costs. Here is the executive summary of an ACLU report published last April:

Every day in our nation’s schools, children as young as five are charged with “crimes” for everyday misbehavior: throwing a paper airplanekicking a trashcan, and wearing sagging pants. In the 2013-14 school year, the most recent year for which statistics are available, schools reported over 223,000 referrals to law enforcement.

A 13-year-old Texas boy who attempted to pay for school lunch with a $2 bill that turned out to be fake faced prison time on charges of felony forgery. In Virginia, a middle school student was charged with assault and battery with a weapon for throwing a baby carrot at her teacher. The criminalization of typical youth behavior has engendered a bizarre reality — students are arrested in schools, places meant to provide safe haven, for behavior that is noncriminal in any other venue . . .

Days after the Columbine High School shooting in 1999, President Clinton cited the first round of COPS grants as a response that would allow schools and police to form partnerships focused on “school crime, drug use, and discipline problems.” In 1998 and 1999, “COPS awarded 275 jurisdictions more than $30 million for law enforcement to partner with school entities to address crime and disorder in and around schools.”
In a nod to the educational mission of schools, lawmakers also asserted that school police would use tactics other than arrests and use of force. Senator Campbell stated that police in schools “would develop or expand community justice initiatives” and “train students in conflict resolution,” a role Senator Lincoln Chafee, a co-sponsor, described as the “most important[]” objective of school resource officers. Members of the House similarly emphasized restorative justice goals and the prevention of police and court involvement. However, as with earlier iterations, the promise of positive support services eased the way for the expansion of policing powers, but the services never materialized. Instead, police, who were neither trained nor certified in counseling or social work, carried on with traditional policing models, addressing perceived rowdiness and disorder through arrests and surveillance of schoolchildren.

The report notes that in 1975, just 1 percent of U.S. schools had a law enforcement presence. Thanks to federal grants driven largely by reaction to school shootings, by 2004 that figure was 36 percent. Today, it’s 24 percent of middle schools and 42 percent of high schools. The report then details how when schools are staffed with cops, administrators grow increasingly likely to defer to SROs, who take a law enforcement approach instead of disciplining kids with punishments such as detention or in-school suspension.

For example, the San Bernardino City Unified School District, in California, makes more juvenile arrests than do municipal police in some of California’s largest cities, and 91 percent of these arrests are for misdemeanors like disorderly conduct. In the Jefferson Parish Public School System, the largest in Louisiana, the Southern Poverty Law Center found that the most common cause of student arrests was “interference with an educational facility.” These findings are consistent with American Bar Association assessments of the juvenile justice systems in many states; the assessments found that school-based referrals and arrests had increased dramatically by the mid-2000s, with schools using the juvenile justice system as a “‘dumping ground’ for youth with special needs.” In one North Carolina county, a full “two-thirds of delinquency case complaints came from the public school system,” and across the state, “[c]hildren as young as six and seven are referred to court for issues that seem clearly to relate to special education status.” Similarly, reviewers in Maryland found that “in interviews, many law enforcement officials across several counties reported a spike in juvenile arrests during the school year due to the presence of school resource officers.”

A 2015 study published in the Washington University Law Review came to similar conclusions. From the executive summary:

Drawing on recent restricted data from the US Department of Education, this Article presents an original empirical analysis revealing that a police officer’s regular presence at a school is predictive of greater odds that school officials refer students to law enforcement for committing various offenses, including these lower-level offenses. This trend holds true even after controlling for: (1) state statutes that require schools to report certain incidents to law enforcement; (2) general levels of criminal activity and disorder that occur at schools; (3) neighborhood crime; and (4) other demographic variables. The consequences of involving students in the criminal justice system are severe, especially for students of color, and may negatively affect the trajectory of students’ lives. Therefore, lawmakers and school officials should consider alternative methods to create safer learning environments.

And here’s a summary of the findings of a 2011 study published in Justice Quarterly:

The use of police in schools has increased dramatically in the past 12 years, largely due to increases in US Department of Justice funding. This study used data from the School Survey on Crime and Safety to assess the extent to which the addition of police in schools is associated with changes in levels of school crime and schools responses to crime. We found that as schools increase their use of police, they record more crimes involving weapon and drugs and report a higher percentage of their non-serious violent crimes to law enforcement. The possibility that placement of police officers in schools increases referrals to law enforcement for crimes of a less serious nature and increases recording of weapon and drug offenses requires that more rigorous research be carried out to assess more carefully the school climate and school safety outcomes related to this popular and costly practice.

And this, from a 2009 study published in the Journal of Criminal Justice:

The high number of disorderly conduct incidences at SRO schools compared to non-SRO schools was consistent with the belief that SROs contribute to criminalizing student behavior. Having an SRO at school significantly increased the rate of arrests for this charge by over 100 percent even when controlling for school poverty. Given that disorderly conduct was the most common charge in this study, these results have serious implications for schools, law enforcement agencies, and juvenile courts. Clearly, disorderly conduct is the most subjective, situational, and circumstantial of the charges studied here. Compared to more objective situations like finding a youth in possession of a knife or narcotics, the decision to interpret disruptive behavior as criminal is done at the officer’s discretion.

And then there are the use of force issues. Cops are authorized to use force in a way that teachers and school administrators are not. This is because cops are trained to use force. Administrators are trained to counsel, de-escalate and discipline in other ways. If administrators increasingly turn to on-site police officers to discipline students, that means more kids will be handcuffed, Tased and beaten. There have been a number of incidents over the years that made national headlines, particularly once video cameras in cellphones became common. Though SROs exist in every state, only 12 states require specialized training for officers who are assigned to schools.

It’s difficult to say whether these incidents are merely anecdotal or if SROs really do use force more often than they should. For one, I haven’t been able to find any data on how often SROs use force nationwide, but even if such data existed, it would be difficult to evaluate — to what would we compare those figures?

However, there are studies showing that SROs use force disproportionately on different populations of students. From a 2015 piece in the Atlantic:

Black students were 16 percent of the total student enrollment in the 2011-12 school year but 27 percent of students referred to law enforcement and 31 percent of students involved in a school-related arrest, according to the U.S. Department of Education Office for Civil Rights data.

Students with disabilities represented about 12 percent of the total student population but accounted for a quarter of those arrested and referred to law enforcement, 75 percent of those who were physically restrained at school and 58 percent of those placed in seclusion or involuntary confinement.

Those numbers are worrying.

I’ll end this post with a quick description of the image at the top. It’s from a t-shirt sold by an SRO union in Northern California. The head of the police union later apologized, but the mentality that would cause such a shirt to be printed and sold in the first place is revealing. Kids require patience, subtlety and empathy. Police officers on average get 120 hours of training in using force for every 8 hours in conflict resolution. If  we know that SROs don’t get much additional training than that, and if we know that the presence of SROs means school officials are increasingly likely to defer to on-site officers for discipline, that’s a huge problem. The criminal-justice system has all the subtlety and nuance of a hammer. If you put that system in charge of school kids’ day-to-day safety, you shouldn’t be surprised if it starts to treat the kids like nails.