That’s what is called an inverted-U curve. Inverted-U curves are hard to understand. They almost never fail to take us by surprise, and one of the reasons we are so often confused about advantages and disadvantages is that we forget when we are operating in a U-shaped world.
Which brings us back to the puzzle of class size: What if the relationship between the number of children in a classroom and academic performance is not this:
or even this:
What if it’s this?
The principal of Shepaug Valley Middle School is a woman named Teresa DeBrito. In her five-year tenure at the school, she has watched the incoming class dwindle year by year. To a parent, that might seem like good news. But when she thought about it, she had that last curve in mind. “In a few years we’re going to have fewer than fifty kids for the whole grade coming up from elementary school,” she said. She was dreading it: “We’re going to struggle.”
***
Inverted-U curves have three parts, and each part follows a different logic. There’s the left side, where doing more or having more makes things better. There’s the flat middle, where doing more doesn’t make much of a difference. And there’s the right side, where doing more or having more makes things worse.
If you think about the class-size puzzle this way, then what seems baffling starts to make a little more sense. The number of students in a class is like the amount of money a parent has. It all depends on where you are on the curve. Israel, for example, has historically had quite large elementary school classes. The country’s educational system uses the “Maimonides Rule,” named after the twelfth-century rabbi who decreed that classes should not exceed forty children. That means elementary school classes can often have as many as thirty-eight or thirty-nine students. Where there are forty students in a grade, though, the same school could suddenly have two classes of twenty. If you do a Hoxby-style analysis and compare the academic performance of one of those big classes with a class of twenty, the small class will do better. That shouldn’t be surprising. Thirty-six or thirty-seven students is a lot for any teacher to handle. Israel is on the left side of the inverted-U curve.
Now think back to Connecticut. In the schools Hoxby looked at, most of the variation was between class sizes in the mid- to low twenties and those in the high teens. When Hoxby says that her study found nothing, what she means is that she could find no real benefit to making classes smaller in that medium range. Somewhere between Israel and Connecticut, in other words, the effects of class size move along the curve to the flat middle—where adding resources to the classroom stops translating into a better experience for children.
Why isn’t there much of a difference between a class of twenty-five students and a class of eighteen students? There’s no question that the latter is easier for the teacher: fewer papers to grade, fewer children to know and follow. But a smaller classroom translates to a better outcome only if teachers change their teaching style when given a lower workload. And what the evidence suggests is that in this midrange, teachers don’t necessarily do that. They just work less. This is only human nature. Imagine that you are a doctor and you suddenly learn that you’ll see twenty patients on a Friday afternoon instead of twenty-five, while getting paid the same. Would you respond by spending more time with each patient? Or would you simply leave at six-thirty instead of seven-thirty and have dinner with your kids?
Now for the crucial question. Can a class be too small, the same way a parent can make too much money? I polled a large number of teachers in the United States and Canada and asked them that question, and teacher after teacher agreed that it can.
Here’s a typical response:
My perfect number is eighteen: that’s enough bodies in the room that no one person needs to feel vulnerable, but everyone can feel important. Eighteen divides handily into groups of two or three or six—all varying degrees of intimacy in and of themselves. With eighteen students, I can always get to each one of them when I need to. Twenty-four is my second favorite number—the extra six bodies make it even more likely that there will be a dissident among them, a rebel or two to challenge the status quo. But the trade-off with twenty-four is that it verges on having the energetic mass of an audience instead of a team. Add six more of them to hit thirty bodies and we’ve weakened the energetic connections so far that even the most charismatic of teachers can’t maintain the magic all the time.
And what about the other direction? Drop down six from the perfect eighteen bodies and we have the Last Supper. And that’s the problem. Twelve is small enough to fit around the holiday dinner table—too intimate for many high schoolers to protect their autonomy on the days they need to, and too easily dominated by the bombast or bully, either of whom could be the teacher herself. By the time we shrink to six bodies, there is no place to hide at all, and not enough diversity in thought and experience to add the richness that can come from numbers.
The small class is, in other words, potentially as difficult for a teacher to manage as the very large class. In one case, the problem is the number of potential interactions to manage. In the other case, it is the intensity of the potential interactions. As another teacher memorably put it, when a class gets too small, the students start acting “like siblings in the backseat of a car. There is simply no way for the cantankerous kids to get away from one another.”
Here’s another comment from a high school teacher. He had recently had a class of thirty-two and hated it. “When I face a class that large, the first thought that I have is ‘Damn it, every time I collect something to mark, I am going to spend hours of time here at the school when I could be with my own kids.’” But he didn’t want to teach a class of fewer than twenty either:
The life source of any class is discussion, and that tends to need a certain critical mass to get going. I teach classes right now with students who simply don’t discuss anything, and it is brutal at times. If the numbers get too low, discussion suffers. That seems counterintuitive because I would think that the quiet kids who would hesitate to speak in a class of thirty-two would do so more readily in a class of sixteen. But that hasn’t really been my experience. The quiet ones tend to be quiet regardless. And if the class is too small, among the speakers, you don’t have enough breadth of opinion perhaps to get things really going. There is also something hard to pin down about energy level. A very small group tends to lack the sort of energy that comes from the friction between people.
And a really, really small class? Beware.
I had a class of nine students in grade-twelve Academic French. Sounds like a dream, doesn’t it? It was a nightmare! You can’t get any kind of conversation or discussion going in the target language. It’s difficult to play games to reinforce vocabulary, grammar skills, et cetera. The momentum just isn’t there.
The economist Jesse Levin has done some fascinating work along these same lines, looking at Dutch schoolchildren. He counted how many peers children had in their class—that is, students at a similar level of academic ability—and found that the number of peers had a surprising correlation with academic performance, particularly for struggling students. In other words, if you are a student—particularly a poor student—what you need is to have people around you asking the same questions, wrestling with the same issues, and worrying about the same things as you are, so that you feel a little less isolated and a little more normal.
This is the problem with really small classes, Levin argues. When there are too few students in a room, the chances that children are surrounded by a critical mass of other people like them start to get really low. Taken too far, Levin says, class-size reduction “steals away the peers that struggling students learn from.”
Can you see why Teresa DeBrito was so worried about Shepaug Valley? She is the principal of a middle school, teaching children at precisely the age when they begin to make the difficult transition to adolescence. They are awkward and self-conscious and anxious about seeming too smart. Getting them to engage, to move beyond simple question-and-answer sessions with their teacher, she said, can be “like pulling teeth.” She wanted lots of interesting and diverse voices in her classrooms, and the kind of excitement that comes from a critical mass of students grappling with the same problem. How do you do that in a half-empty room? “The more students you have,” she continued, “the more variety you can have in those discussions. If it’s too small with kids this age, it’s like they have a muzzle on.” She didn’t say it, but you could imagine her thinking that if someone went and built a massive subdivision on the gently rolling meadow next to the school, she wouldn’t be that unhappy.
***
A half-hour drive up the road from Shepaug Valley, in the town of Lakeville, Connecticut, is a school called Hotchkiss. It is considered one of the premier private boarding schools in the United States. Tuition is almost $50,000 a year. The school has two lakes, two hockey rinks, four telescopes, a golf course, and twelve pianos. And not just any pianos, but, as the school takes pains to point out, Steinway pianos, the most prestigious piano money can buy.6 Hotchkiss is the kind of place that spares no expense in the education of its students. The school’s average class size? Twelve students. The same condition that Teresa DeBrito dreads, Hotchkiss—just up the road—advertises as its greatest asset. “[Our] learning environment,” the school proudly declares, “is intimate, interactive, and inclusive.”
Why does a school like Hotchkiss do something that so plainly makes its students worse off? One answer is that the school isn’t thinking of its students. It is thinking of the parents of its students, who see things like golf courses and Steinway pianos and small classes as evidence that their $50,000 is well spent. But the better answer is that Hotchkiss has simply fallen into the trap that wealthy people and wealthy institutions and wealthy countries—all Goliaths—too often fall into: the school assumes that the kinds of things that wealth can buy always translate into real-world advantages. They don’t, of course. That’s the lesson of the inverted-U curve. It is good to be bigger and stronger than your opponent. It is not so good to be so big and strong that you are a sitting duck for a rock fired at 150 miles per hour. Goliath didn’t get what he wanted, because he was too big. The man from Hollywood was not the parent he wanted to be, because he was too rich. Hotchkiss is not the school it wants to be, because its classes are too small. We all assume that being bigger and stronger and richer is always in our best interest.
***
The definitive analysis of the many hundreds of class-size studies was done by the educational economist Eric Hanushek, The Evidence on Class Size. Hanushek says, “Probably no aspect of schools has been studied as much as class size. This work has been going on for years, and there is no reason to believe that there is any consistent relationship with achievement.”
***
The logic of the inverted-U curve is that the same strategies that work really well at first stop working past a certain point, and that’s exactly what many criminologists argue happens with punishment.