It’s a reality that’s rattled the education world for years: Black and Latino students are far less likely than their white and Asian peers to be assigned to gifted-and-talented programs. The odds of getting assigned to such programs are 66 percent lower for black students and 47 percent lower for Latino students than they are for their white counterparts.

Given the well-known racial disparities in academic test scores that generally determine enrollment in these programs, the gap may seem inevitable. But even among students with high scores on math and reading assessments, black children are severely underrepresented in gifted programs; a high-achieving white student is twice as likely as an equally high-achieving black student to get assigned to such a program. (Interestingly, the gap between Latino and white students virtually disappears when controlling for test scores and other classroom and socioeconomic variables.) According to a new study by Jason Grissom, an education-policy professor at Vanderbilt University, and Chris Redding, a doctoral candidate in educational leadership and policy at Vanderbilt, this suggests that “factors beyond observable student background characteristics are responsible for explaining the Black-White gap in gifted assignment.”

The study, which was published Tuesday in the American Educational Research Association’s peer-reviewed journal, looks at a nationally representative group of more than 10,000 students who started kindergarten in 1998, tracking them every few years throughout elementary school. Controlling for a range of variables—from students’ academic performance to their socioeconomic status to their age—it aims to provide new insight into why high-performing black students are so underrepresented in gifted-and-talented programs. It turns out that the characteristics and tendencies of teachers, according to the study, could be major factors. Black students are three times less likely to be assigned to gifted-and-talented reading courses when those students are taught by non-black teachers versus black ones, the study finds. The authors argue that preventable practices and implicit biases likely contribute to this discrepancy.

In recent years, teachers have been given greater discretion in whom to refer to gifted-and-talented classes. The researchers have warned against drawing the conclusion that non-black teachers are biased against black students, and their study isn’t definitive about what’s causing the underrepresentation. They do, however, cite a number of hypotheses, including that “racialized teacher perceptions” may in part explain why educators interpret their students’ behaviors and abilities in inconsistent ways. “What a teacher may attribute to precocity for one student may be considered disruptive behavior for another,” they write. Conversely, teachers of color may recommend minority students for gifted education at higher rates. Or black teachers may simply be more effective in both motivating black students to improve their own performance and engaging with parents, who are often instrumental in getting their children screened for and enrolled in gifted-and-talented instruction.

The findings are concerning given how few classrooms are staffed with teachers of color, a problem schools have failed to address despite the increasing racial diversity of America’s schoolchildren. In fact, as The Atlantic’s Adrienne Green reported last September, a study of nine major U.S. cities including Chicago, Los Angeles, and Philadelphia found that the disparity between teachers and students of color actually increased in each of the districts between 2002 and 2012. Overall, roughly 80 percent of U.S. public-school teachers are white, and, according to the study, roughly the same percentage of black elementary-school children are taught by teachers of another race. (The vast majority of the teachers studied by the Vanderbilt researchers—91 percent—are white.) “Greater teacher diversity may help ameliorate racial gaps in student assignment to gifted programs,” Grissom and Redding write.

Still, the researchers, who caution against attributing the gaps strictly to teacher bias, conclude that the gifted-and-talented imbalance can be addressed even without diversifying the teaching force. Schools, they suggest, can improve training for teachers tasked with identifying gifted students, explicitly encourage them to address tacit biases, raise awareness about prejudices or stereotypes that lead to inconsistent practices, or adopt universal screening for gifted children. Or all of the above.

Ultimately, the racial disparities in gifted education can widen longer-term gaps in opportunity. Participation in gifted-and-talented programs has been linked with positive future outcomes, including improved academic performance, motivation, and classroom engagement. “The lower likelihood of assignment for high-achieving Black students in classrooms with non-Black teachers,” the researchers write, “diverts gifted services from the very students who may benefit the most from such programs.”