It doesn’t matter where your kid goes to school

By Felix Salmon
August 17, 2011
paper from Atila Abdulkadiroglu, Joshua Angrist, and Parag Pathak, is causing quite a stir in the blogosphere today. (An ungated version is here; see commentary from Mathews, Yglesias, and Salam.)

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

“The Elite Illusion”, a new paper from Atila Abdulkadiroglu, Joshua Angrist, and Parag Pathak, is causing quite a stir in the blogosphere today. (An ungated version is here; see commentary from Mathews, Yglesias, and Salam.)

The main lesson of the paper — which doesn’t surprise me in the slightest — is that students at highly-selective schools, like Stuyvesant in New York or Boston Latin, don’t seem to perform any better than students who might well have gone to those schools but didn’t. In other words, the outperformance of such schools on tests is a function of how selective they are; it’s not a function of how good the teaching is.

A couple of points are worth making here. First is the difference between selective public exam schools, on the one hand, and selective private schools, on the other. The latter are much more expensive, both for parents and on a cost-per-student basis. Reihan Salam is also right that when it comes to hiring teachers, public schools have lots of bureaucratic nightmares that private schools can often avoid.

That said, the lesson of this paper is emphatically not that you’d be better off sending your kid to a private school than to a public exam school. You wouldn’t — as the exam results alone indicate. There aren’t a lot of studies of public vs private schools, but the ones which do exist generally show no difference at all in educational outcome, once you control for the socio-economic status of the kids being admitted. Essentially: middle-class kids who grow up with two well-educated parents and lots of books around the house will generally do very well in school no matter where they go.

This new paper shows that there’s not much difference between public schools, either — even though everybody “knows” that the highly-selective exam schools are better than the alternatives, and many parents will only send their kids to a public school if they get in to one of the very selective ones.

That said, Salam — who went to Stuyvesant — does make one good point which undercuts slightly the thrust of the paper:

The 100 most competitive students at Stuyvesant were indeed very competitive, with almost all attending selective schools, etc. The least competitive students fared less well… I was one of them, incidentally, and I clawed my way out with the help of a few people who continue to be my closest friends, all of whom have gone on to distinguished careers.*

He then quotes Jesse Anttila-Hughes, who teases out the implications of the study, which looked at the kids near the cut-off (she’s he’s one of them), not the brightest kids in the school. The study is essentially taking the experience of kids who are in the bottom 10% of Stuyvesant, and comparing them to kids who are in the top 10% elsewhere. If you’ve ever been a kid, you’ll know that massive difference in relative ability will have enormous repercussions when it comes to your educational outcomes.

Anttila-Hughes continues:

A major reason specialized schools exist is not to help marginal kids do better but to allow superstar kids to do extraordinarily well. Stuy is famously referred to as a “haven for nerds” and like many top schools succeeds by virtue of giving driven and talented kids the chance to do what they want and the resources to do so… I strongly suspect that the causal effect of going to the school is hugely nonlinear in ability.

This might be true — but on the other hand, it might not be. Nerds flourish in the most unlikely places (like Far Rockaway High School), and while there’s a high density of nerds at Stuyvesant, we’d need another study to show how they fare there compared to elsewhere.

In general I think that the obsession over the relative merits of different schools is a classic example of the narcissism of small differences. Some kids fit in much better at this school than at that one — and just as many would be better off at that one rather than this one. There’s no easy way of generalizing, no sense in which School A is in general a significantly “better school” than School B. When it comes to educational outcomes, by far the most important factors determining them are external to the school — the kid’s health, wealth, and home surroundings. And most important of all, of course, is the character and personality of the individual person being educated — something which is much more innate than subject to shaping.

So if your kid doesn’t get in to Stuyvesant, it’s no bad thing — in fact, it might well be a good thing, given that your kid would probably in that case have been near the bottom of the Stuyvesant class. And before you shell out on school fees, ask yourself whether the money wouldn’t be better spent on other forms of education — books, computers, travel, theater, and the like. School’s important. But it’s not nearly as important as most middle-class parents think it is.

*Update: Apologies to Reihan Salam: this is a misleading ellipsis. He was one of the bottom half of Stuyvesant students, not one of the least competitive students in the whole school. He emails to say that “unfortunately, I think it would’ve been far harder for me to get from, say, the tenth percentile to a stronger position than from the 50th percentile”.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

“… I clawed my way out with the help of a few people who continue to be my closest friends, all of whom have gone on to distinguished careers.”

Isn’t this a telling bit? Presumably many if not all of those close friends went to Stuyvesent, and one doesn’t make close friends with elite types who cane help you “claw your way out” if one goes to Dead End High School. Surely much of the point of getting one’s kid into a “good school” is social, for all the meanings of “social” you can name (sometimes it’s frankly “I don’t want my kid hanging out w/the wrong class/race”, sometimes it’s “I want him in an environment that values learning like we do”.)

Posted by jfruh | Report as abusive

Jesse is a man.

Posted by OneEyedMan | Report as abusive

oops! Thanks, fixed

With a lower-class upbringing, I went to a high school that trained steelworkers and housewives (sorry to be sexist, but it was true, 30 years ago). You play the hand that’s dealt you, but you play to win.

Posted by Curmudgeon | Report as abusive

I’m not sure if you’re intentionally being dense and ignoring this issue, but the education one actually gets is only a small part of the reason to attend an elite school.

Posted by inboulder | Report as abusive

@inboulder: I’m not sure if you’re intentionally being rude and ignoring social mores, but there was probably a less abrasive way of saying that.

Posted by strawman | Report as abusive

There’s surprisingly little research on outcomes. Here’s a bit:

1. Work on college / university outcomes says that kids who get into a higher ranked, more prestigious school but who go to a lesser school, do as well in salary. This suggests it’s the person that matters, not the name on the degree. The work in this area is not exhaustive.

2. Every one familiar with grad school admissions knows that when testing counts a lot, that favors the “Ivy” kids who test high. They tested high on the SAT so they test high on the LSAT. A kid who tests high can go anywhere and do as well; there is no “Ivy” benefit, just that kids get in who test well. (This testing difference is then used, btw, to justify higher grades at Ivy schools. Kind of perverse, I’d say, to work backwards from ability testing to justify very high grade levels.)

Posted by jomiku | Report as abusive

“Work on college / university outcomes says that kids who get into a higher ranked, more prestigious school but who go to a lesser school, do as well in salary.”

That study doesn’t say what everyone believes that it says. This effect only deals with the average sat score of the school. When you look at the prestige of the school you get a 13% benefit in salary for those who went to a “most” prestigious school over those who got into the “most” prestigious school and went to a “highly” prestigious school. Check out starting on page 24: 09.pdf llege-prestige-matters.html

Posted by lemmycaution415 | Report as abusive

This is A RIDICULOUS argument.

1) Success in life is dependent upon manners, ethics, morals, class values, cooperation skills, and relationships as much as upon test scores. (Per research into millionaires)

2) Tests measure two things a) genetics and b) exposure to general knowledge.

3) Private Schools filter. That’s what they do. IT’s good for business.

4) By filtering schools can more easily create manners, ethics, morals, class values. They cannot create test scores. If we measured values ethics, manners, morals and class values we would have a higher predictor of success than we would by measuring general knowledge. No one can ‘see’ your scores. They can see your ethics.

5) If we measure by output criteria (what people accomplish) not input criteria (filtering) then small liberal arts colleges win over large state schools and elite schools. (See Sowell)

Education matters. Test scores tell us something. But the truth is that general knowledge, and cooperative skills are more important for most of the area under the curve.

So we are too often having the wrong conversation. It is a conversation driven by the data we have rather than the data we do not have. Which is in itself a common problem in economics and the social sciences.


(PS: Love some of the sarcasm in the comments. Good elitism at play. And we should all defend elitism whenever possible.)

Posted by CurtD59 | Report as abusive

CurtD is correct. The social aspect of schooling is at least as important as the academic aspect in long-term success.

My personal experience as a child in a “good” suburban public school was awful. My academic strengths weren’t challenged (I repeated 6th grade math three times because they had no idea what to do with me after that), and my academic weaknesses weren’t challenged (my handwriting was illegible in 7th grade, and I could not construct a coherent paragraph). Essentially no friends, either in school or outside, until very late in high school — and even now I fear social situations.

I contrast, I have seen my son slowly emerge in a private Montessori setting. He has much the same strengths and weaknesses that I did, but has started exploring friendships in second grade and has developed the manual control necessary for both handwriting and art work.

The funny thing? My test scores were off the charts. It was one of very few things that I did well, so I made a point of doing it very, very well. My son may never test that well — but I see in him the seeds of a more complete adult.

Test scores don’t tell the whole story.

Posted by TFF | Report as abusive

Re: jfruh:

That was precisely my point. Peer groups are important. I also attended a summer program (TASP) that was very influential, and wasn’t connected to Stuyvesant — but of course I was inclined to go because I had a crush on a girl who had gone to the program the year before me, and she was a Stuyvesant student.

Posted by Reihan | Report as abusive

My understanding is that the study considered two metrics: college admissions, and SAT scores. There are easy arguments against both methods.

-Admissions offices try to diversify their student body; a good student at a mediocre high school has a better chance at getting in to a good college than an equally good student at an excellent high school. In effect, there’s a double standard: more is expected from students who attend very good schools.

-The SAT I covers middle school material, so it’s not surprising that students who could have gone to a selective high school did as well on the SAT I as selective high school graduates: this control group was selected, I imagine, by virtue of being at the same level as selective high school students at the beginning of high school (or the end of middle school). It’s a circular proof. The authors of the study would have done better to look at SAT II (subject test) or AP exam results.

I went to Hunter College High School, and every day I am grateful for the education I received there.

Posted by Felsina | Report as abusive

My ex has mild autism or if you prefer, Aspergers. It often comes with ADHD and other learning and social disorders.

He couldn’t read, write or spell at the age of 12. He was considered severely retarded because of his test scores. Fortunately his parents didn’t give up and a very astute and forward thinking Psychologist in Montreal saw he had incredible potential and focused on that. That man became a pioneer in learning disabilities and brain development.

At 13, my ex was able to be integrated back into a normal school and today, he is considered brilliant by his peers. He reads voraciously, sometimes a hefty novel a day. Had his test scores been the only rating, he would have had great difficulty finding employment (now, he heads a department in a large corporation) and certainly never have been able to complete high school.

Although my ex would never be able to achieve the high test scores I did, he has the superior intellect.

@CurtD59, excellent comment!

Posted by hsvkitty | Report as abusive

Wow…very surprising results. I think the key problem with the study is that the tests considered don’t likely show performance at the margins well- the SAT didn’t test anything I learned after ninth grade. One advantage of going to a selective high school is that there are more people at your level which provides flexibility in attending advanced courses… for example, no other school available offered 2 years of courses beyond AP calc or AP computer science. I was able to score 5 on 8 AP exams and get a year of college credit. There were plenty of students to support a variety of intellectual clubs. There was the freedom from the typical American HS sports dominated culture, which is good for many…

Perhaps the downside is that you don’t get to be the best. After years of being the smartest person I knew, it was a bit humbling to be exposed to others on the same level, and those that definitely exceeded me in effort or ability in various courses. I lost a little bit of the leadership opportunity that would have been more easily afforded to me…

Posted by mattmc | Report as abusive

Being a student that was probably in the bottom 10 percent of my class at Stuyvesant, my point of reference is different, but my conclusion the same: Though the vast majority of teachers were bad (ironically, the best teacher I had at Stuy was a Stuy alum himself), it was the student body that made the school special.

I’m sure most of the graduates from Stuyvesant would have fared just as well academically in other high schools (perhaps even better), however, the experience of being surrounded by peers who were just as “nerdy” as you, especially during your teenage years, really helped you develop in ways that this study couldn’t measure.

I had to attend my local high school for summer classes a few times and the difference between the two schools was startling. If I raised my hand to answer a question at my local high school, the kids would immediately ostracize me, which made me reluctant to participate. At Stuy, the same thing would immediately spark a discussion in class and, at worst, lead to a controlled argument. Most teachers just sat back and let the students teach themselves. Even us underachievers would sometimes spend free periods or times spent cutting classes talking about math, science, or politics. I’m sure I wouldn’t have the same luxury in my local high school.

Even as a bottom ten-percenter, I had no problem getting into college and doing well once I got there. If I had gone to my local high school, perhaps I would’ve gotten into a better college since it would’ve been easier to stand out and do well academically. But ultimately, I still managed to be just fine and ended up an engineer.
Again, the value of going to a specialized high school comes down to the student body – even the “bad kids” end up doing okay for themselves.

Posted by Roni1385 | Report as abusive