Bryan Caplan  

Reflections on Gifted Programs

PRINT
Macaulay on Southey... Good News on California School...
Full disclosure: The idea of schools without gifted programs fills me with visceral meritocratic outrage.  In junior high and high school, tracking was the only thing that made my life bearable.  In my memory, normal classes were a combination of Waiting for Godot and Lord of the Flies.  So I thought it best to mellow out a few days before commenting on Arnold's recent posts on gifted programs.

The basic approach of the research Arnold discusses - compare the subsequent test scores of kids just above the gifted threshold to the test scores of kids just below the gifted threshold - seems reasonable to me.  But how should we interpret the results? 

1. The study measures the effect of gifted programs on standardized test scores, not educational attainment, college attendance, college rank, income, or occupational success.  So while the research is good as far as it goes, it doesn't measure the long-run benefits that proponents of gifted programs really hope for.  And if you know the Transfer of Learning literature, you'd shouldn't expect gifted classes to have much effect on test scores unless they teach to the test.

2. You might think that if gifted programs don't boost test scores, they can't boost educational or financial success.  But that's wrong.  In a human capital model, gifted programs could work by boosting non-cognitive skills.  And in a signaling model, gifted programs could work by weeding out and scaring off lower-quality students.

3. At least according to the most knowledgeable person I've talked to, higher-ranked colleges give their graduates a substantially higher rate of return than lower-ranked colleges.  The analogy between gifted and regular classes seems strong enough that we should expect the same result - and be suspicious if we don't find it.

4. Yes, it's easy to object, "The marginal and the average effect are different."  But in the case of gifted programs, the marginal and the average effect probably are different.  I knew many marginally gifted students growing up.  The classes moved too fast for them.  Their choices were: do well in regular classes or poorly in gifted classes.  I can easily believe that gifted classes didn't help their marginal students get better diplomas or better jobs.  But it's hard to believe that gifted classes didn't help their good students get better diplomas and better jobs.

5. The marginal/average distinction is especially relevant when there's censoring.  The highest possible grade is usually an A.  But all A's are not created equal.  An A in a gifted class looks a lot better to selective colleges than an A in a regular class.  If you can earn A's in gifted classes, you benefit: selective colleges will give you a chance.  If you can't earn A's in gifted classes, though, the benefit is harder to see.

One last thought: Some libertarians want government enterprises to run as poorly as possible to expose the evil of the system.  Others want government enterprises to run as well as possible to give taxpayers the maximum value for their money.  When Arnold writes...
Either you believe your bright kids should experience going to class with students who are not so bright, or you don't. If you don't, then pay for private school. G&T allows you to send your kids to private school while claiming they are still in public school.
... he at least sounds like the first kind of libertarian.  Suppose for the sake of argument that gifted classes have zero long-run benefit.  Even so, what's wrong with giving young nerds a classroom of their own to spare them thirteen years of boredom and peer abuse?


Comments and Sharing





COMMENTS (26 to date)

I am always wary of cognitive dissonance. Over on OrgTheory was a link to a HuffPo piece revealing the apparently surprising fact that "most" people feel that university education is too expensive and delivers little value, though "most" college graduates feel that their education was valuable. It is also a factoid that most people think that public education is in trouble but that their local schools are fine.

Like you, I was tracked from the 7th grade - and we separated the merely talented from the truly gifted: AT was not AP. I had to go to summer school to catch up in math to qualify for calculus as a senior, a program to which my brother was already admitted. Cleveland Public had "major work" programs in schools since the 1920s, so for us, this was a program two generations old when we entered.

I still take out learning from those classes in AP American History and PSSC Physics.

I finally gathered all my college credits into one place and one time and completed a BS (2008) and then an MA (2010). Graduating with honors was easy enough. For me, the proof that GT programs are important and valuable for lifelong learning was that in 2007, I got an A in college algebra without a textbook and without a calculator. Gifted and Talented programs paid off for me.

Radford Neal writes:

The basic approach of the research Arnold discusses - compare the subsequent test scores of kids just above the gifted threshold to the test scores of kids just below the gifted threshold - seems reasonable to me.

I think this depends on how gifted programs are run, but for what I take to be the most reasonable way, one would not expect any benefit for the marginal student. Suppose that the content of the programme is designed for the top 5% of students, and they then let in anyone who they believe will do better in such a programme than in the regular class (probably more than the top 5%). If their assessments of this are accurate, performance of the just-admitted or just-not-admitted students should be the same.

fructose writes:
Even so, what's wrong with giving young nerds a classroom of their own to spare them thirteen years of boredom and peer abuse?

I agree with this sentiment, but it should be pointed out that most people in gifted programs are not nerds, and many nerds are not smart enough for gifted programs. I was tracked into gifted programs in junior high, and the average kid was more attractive, and more likely to participate in sports than non-gifted tracks. (Not every class had a gifted version, so I had some average level classes.)

The strange thing was that even the jocks/attractive people in the gifted classes were nicer (and often genuinely friendly) to nerds than the typical average tracked person. (

Jon writes:

The study Arnold is discussing is rather blatantly flawed:

The posterior data presented is norm'd to standard deviations and plotted against the euclidean distance from the cut-off of the proxy. They then compare the intercepts of a linear fit to the two populations (GT and not). This has two oddities.

First, if the underlying trend is non-linear, comparing the intercepts is meaningless. i.e, meaningfully comparing the interceptions implies an assumption that the slope is identical.

Second, the pdf of the normal distribution is a decaying exponential. On one-side of the GT dividing line we see SD scores of 0.6-0.8 and on the other SD scores of 0.8-1.0. The analysis makes the assumption then that a linear fit to two these regions is somehow comparable by its slope and intercept, yet the delaying exponential means that increment in the two categories is very much different (and not the same as is assumed). i.e., even if the slopes appears equally for the purposes of step one, they would actually have different units. Once again, the linearity assumption is falsified.

Lint writes:

From what I remember, normal classes were nap time.

I agree with your criticism with regard to standardized tests. They are not where we would expect to see the benefit, and if normal classes teach more to the test (just an off the cuff assertion, since I'd think honors or gifted students are expected to do well anyway), we would expect zero or negative benefit to be the result.

Also, @Michael, calculators are overrated.

quadrupole writes:

Actually, I am not at all surprised that gifted and talented classes had no impact on the marginal student.

When I was in GT classes growing up, they were required to not be smaller than the standard class size (25 students). At most grade levels, the school could usually marshall at *most* 5 really gifted students. Usually you'd pick up another 5 kids who worked like crazy (and who, in my opinion earned their place). The remaining 15 were generally pretty average to slightly above average kids who's parents thought Johnny or Jane were gifted. These 15 kids weighed down the class significantly.

Which is to say, of course GT classes didn't do much for the marginal student, the marginal student didn't really belong there in the first place.

JimS writes:

As a middle school teacher who teaches an advanced Language Arts section, I would agree with the comments that it's not surprising that the borderline kids don't show a lot of improvement on state standards tests or on the NAEP. I don't spend nearly as much time on basic grammar with that class. The kids who have been correctly placed have it down, and are ready for more advanced writing and lit analysis.
The kids who are there due to squeaky wheel parents aren't being well served with the placement. I try to work on those issues with them, but it is my largest class and there is limited time for remediation.

Steve Sailer writes:

Bryan says:

"I knew many marginally gifted students growing up. The classes moved too fast for them. Their choices were: do well in regular classes or poorly in gifted classes."

Excellent point about why marginal case analysis wouldn't work here.

Daublin writes:

I agree in principle, Bryan, but I found gifted break-out programs to be really unpleasant in practice. I was in one of those for a year in elementary school, one of those were a bus of kids goes off to gifted land once a day every week or two.

All we did in the break-out, though, was go through large piles of really boring stuff. I was so excited at first, but gradually I could not get over the material they were covering. I hated it by the end of the year, and I didn't go back a second year. I found it much more pleasant to snooze through class and read math books.

This is not to say I hate advanced classes in general. AP classes were a godsend in high school, and the few normal-level classes I took seemed terribly slow and boring. I would have loved to have something like that in middle school.

To try and explain the difference, maybe it's because gifted programs don't get especially good teachers; you know, anyone who could really teach that class could get rich doing something else. Another possibility is that the AP classes teach to a test rather than to some fad-ridden output of a local school system. In contrast to many commenters, I find the AP tests quite good.

David E writes:

Excellent post. A lot of kids that really belong in the gifted programs would be completely turned off by education (for the rest of their lives in some cases) if the gifted programs were not available. I've seen it happen.

Noah Yetter writes:

Bryan, if parenting doesn't make any difference to how your kids turn out, how can school? You're being massively inconsistent.

More to the point you're completely ignoring selection bias. Gifted programs don't benefit kids success in life, they only accept the kids that will end up being the most successful. Selective colleges do the same.

Yes, there is the "shelter the nerds" effect, but that hardly merits the money cost, and I say that despite the fact that it personally benefited me.

Please also see my other comments on past Econlog G&T posts. These programs are a sham.

eccdogg writes:

Very interesting.

I found GT classes in elementary school to be a complete waste of time.

However, I thought tracking was was quite good and AP classes were great. I was tracked from 5th grade onward and it was nice to be in a class that went at a faster speed. My grades tended to be better when the class was more challenging.

I do think it is a little hypocritical for educational beaurocrats and politicians to go so crazy about kids of different income levels going to the same school just because it is in the same building. I went to a totally different school than kids in average classes, and low/LD classes. The only interaction I had with those kids was in the lunch room and on the Football/Wrestling team (yes some kids are both Nerds AND Jocks)

However, I guess just being in the same building with Avg/Low kids does help you learn how to get along(deal)with those folks in general society.

Funny story. My JR. High was pretty rough and all the smart kids were tracked together, but for PE two classes were merged together. For some insane reason the PE classes that were merged for us were the GT kids and the Low/LD kids. Far from protecting the Nerds it was like leading lambs to the wolves. I remember we had a week of wrestling in PE that was especially brutal. To this day I can't figure out what the people running that school were thinking.

eccdogg writes:

Another thought.

How does this idea jibe with your general call for educational austerity and means testing.

If we should do less in the public sector so people generally switch from public to private options wouldn't getting rid of G&T be a reasonable first start?

Why should the taxpayers be forced to pay for a premium option that largely benefits upper income and IQ individuals? If we are going to spend a limited amount of money are upper income IQ folks really the most in need of our assistance?

tom writes:

Bryan, the study that set Arnold off says exactly the same thing about the marginal students. It guesses that the work was too hard and that being at the bottom is a vicious cycle.

This is only issue I have seen where Arnold's priors seems to be impairing his ability to think about the subject or to evaluate a paper. Maybe he was attacked by GT geeks in a dark lab decades ago? (That would actually more acceptable than the idea that he'd like to force the best students out of public schools so that their parents won't be as accepting of the public education system and costs.)

Evan writes:

@Noah Yetter

Bryan, if parenting doesn't make any difference to how your kids turn out, how can school? You're being massively inconsistent.

Having read Judith Harris' book that started all this, I can explain why schooling could theoretically affect how kids turn out:
Everyone has different personalities for different situations. The reason parenting doesn't affect how kids turn out is that it only affects our "parent personality," the way we act when our parents are around. Schooling, by contrast, shapes our "peer-group personality," which we use frequently our whole lives. Gifted programs give you better quality peers, encouraging you to develop a better-quality "peer-group personality."

I don't know if this really works in the specific instance of Gifted programs, I'm just extrapolating based on the personality-formation mechanism Harris describes. My main point is to show that Bryan's views are not inconsistent, based on an understanding of child development.

Brad Warbiany writes:

Bryan: "At least according to the most knowledgeable person I've talked to, higher-ranked colleges give their graduates a substantially higher rate of return than lower-ranked colleges."

Students who graduate from elite colleges tend to earn more, yes. But do they actually have a higher rate of return? A recent study suggested no.

Controlling for what schools a student applied to [whether they were admitted or not] equalized everything. A student who applied to Harvard but went to a less prestigious school ends up earning as much as the student who applied to Harvard & attended Harvard.

Foobarista writes:

I was in G&T in late elementary and junior high school, and it was a welcome escape from the drudgery and boredom of "normal" classes. High school didn't have G&T and also didn't have AP (which was just starting then, and which my relatively poor district didn't have), so it was a hell of boredom. Since it was so infernally boring, I did just well enough to pass (doing well in class was a good way to get beat up), went to community college, and transferred to UC Berkeley.

I always figured I could easily have gone straight from 7th grade to CC without missing much of anything academically, but this wouldn't have occurred to my parents (or to me in those days).

Finch writes:

I'm not sure whether, when the original blog posters refer to G&T, they are referring to "tracking" or "enrichment".

Tracking means separating out some fraction of the population so the rest can do stuff the separated folk can't. Drop the under-85 IQ people and the room becomes physically safe and people can sit still and pay attention. Drop under-100 and you can learn stuff. Drop under-115 and you can learn calculus. Drop under-130 and you've got an econ grad school. Drop under-145 and you've got a science grad school. :)

Enrichment means trying to use the cognitive surplus of the smart kids to do something else that's supposed to be interesting without distinguishing them too much from the normals. I suspect it's what the G&T complainers are talking about. I did it, and at the time I wondered why they weren't letting us do something useful with ourselves, like learning more math...

Tracking is a huge win because it let's you accomplish more (potentially with the same resources)(*). Enrichment, I'm not at all sure about - it attacks boredom, but isn't all that effective.

(*) It is not perfectly clear to me that the under-85s are better off without the over-115s in their class, but I kind of think that they are.

Do we want elites who love government programs or elites who loathe government programs?

Dan Weber writes:
Bryan, if parenting doesn't make any difference to how your kids turn out, how can school? You're being massively inconsistent.

Well, he does hedge his bets. Even if there is no increase in outcomes, childhood is generally better when not being persecuted. (And Bryan's thesis is that parents can affect the joy of childhood.) So we might as well put the above-average kids somewhere where they aren't assaulted.

I also had Bryan's reaction to Arnold's suggestion: it tries to make the schools as bad as possible so that people have to leave them.

I switched from a private to a public school expressly for the public school's advanced track, so I'm sure that affects my outlook.

dysgenic writes:

The existence of G&T has strong negative externalities on the marginals who didn't qualify. There is no signal to say that one was a marginal candidate and there may have been some randomness at the margins.

Frances Woolley writes:

"The analogy between gifted and regular classes seems strong enough that we should expect the same result - and be suspicious if we don't find it."

Higher ranked colleges have an impact on earnings in part because of signalling - being able to get into and graduate from a high ranked college is a signal of ability.

On the other hand "I was in a gifted elementary school class and that's still significant enough to earn a place on my resume" is not exactly a signal of ability! (Although - and this ties into Arnold's point - in England going to the *right* school (Eton, Harrow, etc) *does* matter, not only for university admission, but in later life).

Larry writes:

I started high-school in non-gifted English class. My teacher chose to have the class read Capote's In Cold Blood aloud in its entirety. I was so offended that I spent my classtime reading the unabridged Les Miserables and ostentatiously paid no attention to the goings on around me. I changed to gifted English the next semester and never looked back.

Tracy W writes:

Larry - your English teacher was reading Capote's In Cold Blood and *that* offended you?
When I was 16, I got an English teacher with whom the only book we managed to finish all year was a children's short story. On other grounds, I became convinced that that was because the book was at the limits of her intellectual level.

Four of us held a party on our last day of English ever.

Laura Walsh writes:

I think there should be a distinction, especially in elementary and middle school, between full time gifted classes and pull-out programs. When I was in 6th grade, in a regular class, I began getting very low grades. My parents were alarmed and got me into a full time gifted program in a neighboring school district. I felt as though my whole mind was waking up. It was a completely transformative time for me. I am certain I would have had a much different educational trajectory had I not had that experience.

On the other hand, I have taught several gifted pull-out programs and have been marginally unsatisfied with them. Gifted students need REAL content, not "Fun and Games for Smart Kids". Try as I might, the pull-out programs often were by necessity the latter, because I was not allowed to incorporate anything that the regular teachers might want to include in their classes. It was still beneficial for the GT students, especially for social and emotional reasons, but academically, it was more problematic.

My heart is with full time programs and even separate schools.

Laura Walsh writes:

@eccdogg: Why should the taxpayers be forced to pay for a premium option that largely benefits upper income and IQ individuals? If we are going to spend a limited amount of money are upper income IQ folks really the most in need of our assistance?

High IQ doesn't just occur among the well off, though it might appear to be so at times. The really, really smart lower income kids are the worst off. They do not have the resources or support to attend private schools. Often, their only hope is a good program in the public schools. Yes, many private schools have scholarship programs, but they often target the high achiever, good student, not the underachieving rebel or the disappearing dreamer.

Comments for this entry have been closed
Return to top