A recent study examined whether gifted programs benefit students at the margin: those who barely “made the cut” for admission into a program and those who barely missed it. The study found that students in both subsets performed approximately equally on standardized tests a couple years after demarcation.

Obviously, this study says nothing about those students who easily “made the cut”—those who are the most gifted. (Other research indicates that these highest achievers do benefit from being around similarly gifted peers.) Instead, the research only looked at whether gifted programs are beneficial to students at the margin. And the answer is actually a somewhat-counterintuitive maybe: gifted programs might be beneficial for students on both sides of the margin. (I explain this below.)

Subsequently, a couple news outlets reported that the findings of this study proved that gifted education programs were ineffective:

“If the gifted and talented programs are effective, then the marginal students should end up with higher test scores than the marginal students in regular classes. If they’re not effective, then both sets of students would have around the same scores.” The Atlantic

“A new study has shown that gifted and talented programs have no effect on student learning.” Teach for America Blog

Fortunately for gifted programs, these absolute statements are inaccurate. It implies that a lack of difference in scores proves the ineffectiveness of gifted programs. That the study concluded that students on both sides of the cutoff performed equally does not mean that gifted programs are generally ineffective. It doesn’t even mean they’re ineffective for students at the margin. In fact, it might prove the opposite.

Allow me to illustrate with a simple thought experiment. Let’s suppose that, all other things being equal, a student who believes she’s smart and destined for success has an advantage over a student who thinks herself inferior and destined for mediocrity. The study’s researchers recognized this and offered it as an explanation for the score similarities between students on both sides of the margin: Students who barely made the cut probably earn worse grades than their more “gifted” peers, leading to discouragement and “checking out.” Meanwhile, students who barely missed the cutoff likely do better than their peers in their non-gifted classrooms, boosting their confidence and receptiveness to teaching. The researchers call this “self-concept.”

If this theory is correct, then high self-concept gives students an academic boost. This is significant. Cutoffs do have to be made somewhere, and due to the nature of statistics, those on either side of the cutoff should have virtually equal academic ability. So (1) if the kids who barely missed the cut are aided by a self-concept boost, (2) if marginal students in the gifted programs suffer from a self-concept loss, and (3) if both of these groups perform the same on standardized tests, then gifted programs are effective.

In a nutshell, Group 1 benefits from high self-concept, Group 2 suffers from low self-concept, but both groups score the same. So how does Group 2 match Group 1? Well, they enjoy a different benefit: the gifted program. In fact, the gifted-program benefit is great enough that it overcomes the lower self-concept and matches the confidence boost enjoyed by kids who barely missed the cut. In other words, assuming this theory, the cutoff actually helps the kids who barely missed it.

Therefore, the true take-away from the study isn’t “don’t offer gifted classes” or “gifted programs have no effect on student learning.” It’s more a message to individual students and their parents: if you are right at the cut-off, you might do just as well in the “regular” class as the gifted class. That’s worth knowing. And that’s what should’ve been reported.

Gifted programs benefit the highest achievers. And as this study demonstrates, they might benefit the students who barely gained admission—and even the students who barely missed the cut. Maybe everybody wins.

Item Type: