In Defense of Suspensions
Note: Our guest-blogger this week will be Max Eden, a senior fellow of education policy at the Manhattan Institute.
That school suspensions are unconscionably high, and that we have a moral imperative to do something big in response has gone beyond mere conventional wisdom: It is well-nigh unquestionable, something that "no one doubts." It is precisely when a claim reaches this status that it must be questioned firmly, with clear eyes.
The case for an aggressive approach to reducing suspensions rests on three assertions: (1) The disparate impact of school suspensions is evidence that they are racially motivated; (2) Suspensions do significant harm to students; (3) "Restorative justice" is a viable and more humane alternative, so we can reduce suspensions safely.
If we accept these three assertions as entirely true, then reducing suspensions is a moral imperative rather than a matter of policy tradeoffs. But these claims are only true up to a point, and accepting them without qualification can lead to tradeoffs that may do much greater harm than good to students and schools. Let's examine each assertion in turn:
The disparity in suspensions is due to racial bias. There is a dramatic difference in the rate of suspensions by race. According to one study, 1 in 6 African-American students were suspended at least once, compared to 1 in 14 Latinos, 1 in 20 Whites, and 1 in 50 Asians. The University of Pennsylvania's Shaun Harper conducted a major study of suspensions in southern states that showed some disparities far too striking to be explicable without racial bias. Another study showed that white teachers tend to view black student behavior more negatively than black teachers. So evidence supports the claim that the disparity in suspension rates is attributable to racial bias—up to a point.
The question is—just how much of the phenomena can be explained by it? 100 percent? 75? 50? 30? Studies like these can point us to a problem, but they can't truly take us inside the issue. If we really think that, say, 75 percent of the disparity is due to teachers' racial bias, then it would make sense to invest a lot in professional development to re-educate educators.
But if we pursue wholesale policy changes on the assumption racial bias is solely responsible for the disparity (the assumption behind the Department of Education's "Dear Colleague" guidance on discipline) then we run a strong risk of overcorrecting. Rapid overhauls of the entire system might well breed rampant disorder in schools—and in places where that approach has been taken, the results aren't looking pretty (but more on that in a moment).
Suspensions do great harm to students. Another oft-heard claim is that school suspensions place students in the "school-to-prison pipeline." One landmark study out of UCLA showed that students who were suspended were dramatically more likely than their matched-peers to drop out, commit crimes, and take public assistance. This study was rigorous, matching students across several observable variables, and should be taken seriously—up to a point.
While a study like this can prove that suspensions are associated with those outcomes, it can't truly support the claim that suspensions caused these outcomes. There are differences between, say, a kid who punches someone else and one who doesn't, which controlling for race, class, and family composition doesn't account for. You would rather expect long-term differences between a troublemaker and a well-behaved student of a similar background; you wouldn't necessarily conclude that the suspension caused the differences. So, policymakers should be extremely wary of making any claim of causality from a study like this.
There are also two sides to this ledger: while claims of the benefits of not suspending disruptive students are likely oversold, the harm done to students by disruptive peers is certainly underappreciated. A study by the National Bureau of Economic Research found that "exposure to a disruptive peer in classes of 25 during elementary school reduces earnings at age 26 by 3 to 4 percent ... [and] removing one disruptive peer from a classroom for one year would raise the present discounted value of classmates' future earnings by $100,000."
It's fair to question whether suspensions are truly the best tool to maintain classroom order. But it's important to recognize that any major policy change is likely to have tradeoffs, harming some students even as it helps others.
We can reduce suspensions safely. Of the three claims, this is the most dubious. There certainly are case studies of schools that have successfully adopted a "restorative justice" model. And I have no doubt that it can work—again, up to a point.
Unfortunately, that point doesn't take us very far. Much of the reliable evidence on the effects of rapid, large-scale school discipline reform in major urban districts is pretty grim.
In Chicago, where a thorough study of the effects of shortening suspension length found a significant worsening of student-reported peer-relations, and teacher-reported crime and disorder. In New York City, Mayor Bill de Blasio has also undertaken a major suspension-reduction initiative. In de Blasio's first year, according to the NY State Education Department, the number of violent incidents in schools increased from 12,978 to 15,934, the steepest increase on record.
These aren't the only cities where things are getting worse, but we don't have great data on many other cities. We don't, for example, have great data on St. Paul, where superintendent Valeria Silva was recently ousted over her discipline reforms. In an incisive postmortem, the Center for the American Experiment's Katherine Kersten quotes St. Paul Police spokesman Steve Linders saying that fights that "might have been between two individuals ... [now become] melees involving 40 or 50 people." Kersten also relates the story of a teacher who, after being crushed into a shelf by a student, asks her students to use a secret knock before she'll open the door to her classroom. Denise Rodriguez, president of the St. Paul Federation of Teachers, demanded, "Do students and staff deserve to come to work every day and not expect to be assaulted?"
Maybe not, given the way things are headed. The Department of Education might well have applauded St. Paul for following its "Dear Colleague" guidance by trying to narrow the racial suspension gap, and no one could have made a data-driven case against it. Things had to get truly dystopian before they could get better.
We simply don't have enough data to evaluate the effects of discipline changes in most American school districts. That means that the policymakers pushing suspension-reduction reforms are doing so quite literally ignorant of the consequences of their actions for poor and minority students who are just trying to behave, learn, and have a fair opportunity at life.
Critics of school suspensions may be onto something, and there's no doubt that their hearts are in the right place. But by pushing reckless, large-scale, administrative overhauls of school discipline, reformers may well be hurting the kids they're trying to help.