Tracking, the practice of putting a small group of higher achieving students into separate advanced or honors classes, isn’t popular with progressive educators. Previous research has pointed out that it exacerbates inequality in our schools because higher income and white or Asian kids are more likely to get tracked into the elite classrooms. Students who aren’t chosen can become demoralized, or the curriculum in the average class can get too watered down. Great teachers and extra resources get steered to these honors programs, leaving the kids who need the most help with less. Researchers have sometimes found that lower-achieving kids are worse off in schools that track.

Now two fresh studies, both published in March 2016, make a compelling case for continuing to cream off the top students and teach them in separate classrooms. One, from the Brookings Institution, suggests that the United States won’t produce as many students, including blacks and Hispanics, who can master higher mathematics if schools don’t begin preparing them separately, starting in eighth grade. The second one, from two economists, finds that tracking can close the achievement gaps both between high-IQ blacks and whites and between high-IQ Hispanics and whites.

The Brookings researcher, Tom Loveless, found that states that track more students into different ability levels in eighth-grade math wind up with more students scoring better on Advanced Placement exams, typically taken by top students during the senior year of high school. States where tracking isn’t practiced as much had fewer students hitting a passing score of 3 or higher on AP tests.*

“We’re talking about a very rarefied group of high-achieving kids who are taking the toughest courses and the toughest tests,” said Loveless, the author of the study. “My point is that they don’t just get there out of thin air. You need to cultivate talent over time in mathematics.”

“I draw the analogy to sports,” he added. “When we hear about the high school starting quarterback who’s a great star, he started playing football when he was eight. We’re not shocked to hear that these kids were identified very young, and that they were offered completely differentiated opportunities to cultivate their talent.”

Math isn’t football, of course, and schools strive to help all children excel at math. But Loveless’s research raises an age-old question in education of whether excellence is sacrificed by well-intended efforts to promote equity.

Tracking in eighth-grade math—steering some to algebra and most others to another year of general math—remains popular across the United States. It’s also a critical decision in a student’s life. Kids who don’t study algebra in eighth grade proceed on a path that effectively shuts them off from calculus and advanced science classes. On average, Loveless found that states tracked about three-quarters of eighth-graders in math, with Arkansas tracking the least (50 percent) and Nevada the most (97 percent). (The data came from surveys of school principals conducted by the National Assessment for Educational Progress, or NAEP, in 2009).

Four years later, in 2013, roughly around the time that these eighth-graders would have been eligible to take an AP exam, Loveless found that states with more tracking had more passing scores. For example, Utah tracks 89 percent of its students in eighth-grade math, and 70 percent of AP test takers in Utah scored a 3 or higher. In Texas, only 57 percent of students are tracked, and only 52 percent of the state’s AP test takers passed.

Loveless was concerned that states with more poverty might be producing fewer high AP scores, and that his results might be unrelated to tracking. But he controlled for poverty, and still found that a tight relationship between tracking and AP scores. The relationship also held true for minority groups. Higher percentages of black and Hispanic students scored well on the AP test in states where there was more tracking.

The other study found big benefits for high-achieving minority students who were tracked into “gifted” classes. It’s a working paper at the National Bureau of Economic Research, written by David Card at the University of California, Berkeley, and Laura Giuliano, at the University of Miami. They studied an unnamed school district, described as one of the largest in the country, where elementary schools were required to establish gifted classes even if there was only one child who scored high enough to qualify. So the rest of the seats would be filled by high-achieving kids in that school who had missed the “giftedness” cutoff. Because the schools are quite segregated in this district, many high-achieving black and Hispanic children ended up in these elite classrooms.

The economists found that these high-achieving black and Hispanic students flourished. Their academic gains were equivalent to what researchers are finding at the best charter schools, and the benefits persisted through at least sixth grade.

As importantly, there was no trade-off between fostering excellence and promoting equity in this case: the researchers didn’t find negative consequences for students who weren’t selected or narrowly missed the cutoff to enter the gifted classrooms.

Interestingly, the benefits to joining these classrooms were the biggest for minority students. The researchers didn’t find great gains for high-achieving white students. Whites of similar academic abilities did about as well in the regular class as they did in the gifted class.

“We show that minority students have lower achievement scores than white students with the same cognitive ability, and that placement in a [gifted] class effectively closes this minority underachievement gap,” the authors wrote.

Superior teaching in the gifted classes couldn’t explain this, they said. So Card and Giuliano hypothesize that two other things are going on: teacher expectations and peer pressure. In regular classrooms, teachers may be overlooking higher-ability minority students and not pushing them as hard as they could be pushed. But in gifted classrooms, teachers are expecting excellence from everyone. Secondly, the researchers wonder if smart minority students are particularly susceptible to peer pressure in regular classrooms, where it’s not “cool” to be smart. In the gifted classrooms, classmates may be more supportive of working hard and getting A’s.

The big argument against tracking is that black and Hispanic students are penalized by it. But perhaps tracking is what is needed to get more blacks and Hispanics into the elite ranks of top scientists and mathematicians.

This post appears courtesy of The Hechinger Report.

* The original version of this story published by ​The Hechinger Report​ and The Atlantic incorrectly characterized the AP test scores that were compared in the Brookings study. We regret the error.