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?