Basically I’m just going to focus on the political economy, and problems of information asymmetry and game theoretic issues with the assertion that armed guards in schools will lead to no more mass shootings. For those who missed Wayne LaPierre’s speech, Jason Linkins at the Huffington Post provides a really good assessment of it.
With that said, who would pay for all these guards? LaPierre seemed to intimate that it would be a public expenditure, so that makes it a political economy issue. It also means that the payment would be made with tax dollars. Putting aside the incongruity of LaPierre and the Republican party’s hatred of taxes with the recommendation that we spend $18 billion tax dollars on a program, how could we finance it? A big tax on firearms sales would probably do the trick. In fact, while financing the protection of schools it would also drive down the sales of guns, making it less expensive to secure schools. Perfect! Any Congressional rep who doesn’t approve can go on Piers Morgan’s show and explain why Grover Norquist’s pledge is more important to them than a tax that would prevent, according to the NRA, another mass shooting.
But this is all academic, because some basic game theory points to the fact that armed guards or more police in schools will not affect mass-shootings on average across cases. There have been a number of articles pointing out the fact that armed guards at Columbine High School were not able to stop the event, and they hint at why but not in a formal way. An inherent problem with the argument that armed guards will stop a shooter is that the argument hinges on perfect information. We have perfect information after the fact, which makes it easy to say “if only there’d been an armed officer there it wouldn’t have happened.” The problem is that we are taking our perfect knowledge of the event, and making the false assumption that the officer or guard would have had the same information. Hence why the school resource officer was not inside Columbine High School when the shooting started.
Realistically, we have to control for two things in a tactical situation. Something called an OODA Loop, and the problem of information asymmetry. OODA stands for “observe, orient, decide, act” and the loop is the cyclical process of someone going through these steps repeatedly as quickly as possible. The standard example of OODA loops is fighter pilots in a dog fight. A fighter pilot has really good information though; they have radar, they know if the other airplane is friendly or not, and if not, they know it’s time to fight. From there it becomes a one-on-one process of who can go through the OODA loop faster to gain a tactical advantage. When someone cannot complete their OODA loop (too much information coming at them or confusing signals), they’re stuck and are at a tactical disadvantage. We see this in the Columbine example; the officers were observing (O) a confusing situation and getting conflicting information, having to constantly re-orient (O) to either deal with shooters or victims, and thus had to make decisions (D) and actions (A) that prevented them from actually dealing with the shooters, who had superior firepower. This is an external problem, since OODA loops are affected by environmental factors; if we were going to make guards in school effective there is also a decision set problem at the individual level.
Prisoner’s Dilemmas are a standard asymmetric information game. The actors must make decisions based on a lack of information about whether to cooperate or defect. If both cooperate all is fine; if both defect the effects on both a minimal. If one cooperates and the other defects, the cooperator suffers greatly. For the sake of a school guard this could mean letting a person pass (cooperate) or making the decision to draw a gun (defect). The problem is: who is a shooter in a public place full of kids and adults? If you pick wrong it’s at the very least a lawsuit, at the worst you fail to intervene because you ‘cooperate’ and are ‘defected’ against (e.g. shooter shoots you first), or are in the wrong place as a shooting happens. And of course, since a shooting has a very short time span, the opportunity to make up for a wrong decision is nil.
The problem with placing a guard in this situation and assuming they would be effective is thus: they cannot complete an OODA loop because there’s too much information, and we mistakenly attribute our perfect after-the-fact knowledge to their completely (dangerously) imperfect pre-event knowledge. Unless the school guard uses a very aggressive strategy, the likes of which would probably not fly legally, he or she would basically be useless as a preventative force on average across cases.
You’ll notice I use the term “on average across cases” a lot. It’s an important term when talking about public policy. Indeed, there will be shootings and violence, and the goal of public policy is not to claim to prevent all bad things from happening everywhere. The goal is to create an environment where the likelihood of an event like a mass shooting is lowered; since there is no evidence that on average across cases, more guns, armed guards, children charging shooters etc. have any effect on the likelihood of a mass shooting I as an empiricist am left with one conclusion. The problem is access to these types of guns themselves.