Some commenters on the thread about the 7-year-old being criminally prosecuted for bringing a toy gun to school linked this to zero tolerance policies, and suggested such policies are as a general matter unsound. I think the question of zero tolerance policies is a bit more complex, so I thought I’d blog a bit about it.
1. To begin with, a zero tolerance policy means zero tolerance for some class of behavior, with some mandatory penalties. This suggests that some such zero tolerance policies might be good, depending on what is banned and what the penalties are. Zero tolerance for, say, fighting (with an exception for self-defense), with mandatory penalties starting with an afternoon of detention and working up from there, might well be sound. Zero tolerance for bringing sharp knives to school, with no exception for honest mistakes, but with mandatory penalties starting with a week of suspension and working up from there, might also be sound (depending on the circumstances). Zero tolerance for bringing toy pistols to school, with the penalty being a call to the parents, admonition to the child, and confiscation of the toy might be sound as well.
Zero tolerance for any of these things with mandatory expulsion would be unsound. Zero tolerance for seven-year-olds’ bringing toy guns to school with mandatory criminal prosecution would certainly be unsound. (That there’s no evidence that there was indeed such a formal zero tolerance policy within law enforcement in the story that triggered the discussion. But the decision to criminally prosecute the seven-year-old for such innocent, harmless behavior suggests that something similar to a zero tolerance policy might be being informally developed.)
2. Of course, one could argue that zero tolerance policies are always bad, because they improperly restrict administrator discretion. Whatever the behavior, the theory goes, it should be up to administrators to make the decision. If the behavior is egregious enough, they’ll punish it every time. But they should be able to look at the facts of each case, including by asking whether there was an honest mistake, whether the student has generally had a perfect disciplinary record, and whether there are other extenuating circumstances. We hire them to make disciplinary decisions, so we should trust them to exercise their discretion in making such decisions.
This is in fact the almost invariable rule with regard to prosecutors’ decisions to criminally prosecute, and one can certainly have such a rule with regard to administrator decisions, too. But there are several problems with such untrammeled discretion, or with rules that have certain kinds of exception, including:
a. Some administrators will fail to properly punish some students, especially if the punishment means extra time, effort, and risk of political difficulties for the administrator.
b. Some parents might rightly or wrongly perceive that their children are being treated unfairly, by being punished when others escape punishment. This can cause a lot of dissension, cost a lot of time and effort, and perhaps even cost a lot of money — especially if the parents (1) claim that their children were discriminated against based on race, religion, sex, and the like, or (2) have a lot of money to hire lawyers (or are lawyers themselves).
c. Some unduly optimistic students might think that administrators will give them a break (perhaps because the students have well-connected parents, or are on the football team, or just because the students think everything will turn out well for them), even if the administrators actually won’t. This means that discretionary rules may lose some of their deterrent force precisely because they are discretionary.
d. Some questions — such as whether a knife was brought to school by mistake — are hard to reliably answer, because of the risk that students will lie, both to protect themselves and to protect their friends.
This suggests that there are substantial possible benefits to a zero tolerance policy, where at least some punishment is required with no possibility of discretionary waiver, and no “honest mistake” defenses. And while there are some costs to such a policy, those costs are comparatively low for modest punishments, though much higher for greater punishments.
This leads me to tentatively think that the problem is not with zero tolerance policies as such; some zero tolerance policies are probably a good idea, depending on the sorts of student behavior problems and administrator-parent tensions that might be present in the school or the school system. The real problem is with zero tolerance policies that (1) are defined too broadly (for instance, to include butter knives, nerf guns, easily identifiable over-the-counter medicine, and the like) or (2) carry punishments that are highly unjust as to a nontrivial fraction of the incidents in which the policy will be applied.