New research challenges a piece of common wisdom about special education: that black students are too often told they have a disability.

It’s true that 15 percent of black students in the U.S. are identified as disabled, while only 13 percent of white students are. Some worry that misplacing black students in special education segregates them and lowers expectations for their success. The disparity has even prompted action from the federal education department, which has long cautioned school districts against over-identifying students of color.

The latest study, published in the peer-reviewed journal Educational Researcher, joins other recent research in calling these concerns into question — and suggests that bias may be at work in the other direction.

“These well-intentioned efforts appear to be targeting the wrong problem typically being experienced by racial and ethnic minority children attending U.S. schools,” write the study’s seven authors.

The researchers controlled for factors like poverty and student test scores to determine whether similar students of different races are identified as disabled at different rates. Their conclusion: kids of color are actually less likely than white children to be identified as having disabilities.

For example, the study finds that about 75 percent of the lowest-achieving white boys in fourth grade are identified as disabled, compared to under 50 percent of fourth grade black boys of the same achievement level.

That finding, of underidentification of students of color, generally holds across different races — black, Hispanic, Native-American, Asian-American — different years and grades, and different disabilities. Asian-American students appear to be especially underidentified for special education.

Paul Morgan, one of the authors of the study and a professor at Penn State, says that some have mistakenly taken the higher raw numbers for some racial groups as evidence of bias.

“You don’t hear a statistic like ‘minority children are twice as likely to have asthma’ and automatically conclude that pediatricians are racially biased, and let’s stop identifying children who are minorities with asthma,” he said.

“I’m all for monitoring for the potential of racial bias, but there’s a correct way and an incorrect way to do that,” Morgan said. The correct way, he says, is to “look at similarly situated or otherwise similar children in the school” — by comparing students with comparable achievement levels, for example. 

The findings are likely to prompt pushback. A 2015 study from Morgan and colleagues reaching similar conclusions received a sharp critical reaction from some other academics. In a response published in the same academic journal, several researchers accused Morgan and colleagues of oversimplifying a complex issue. (Morgan and one co-author wrote a reply to the critique.)

The dispute turns in part on a question about what data set was used in the 2015 study. Morgan says his latest work, which uses different data, is partially meant as a response.

The critics did not address Morgan’s argument about the limitations of past research. They did say, though, that the study goes too far in implying that underidentification of students of color is always a problem and the opposite never is.

There “is not a broad and sweeping witch hunt targeting overrepresentation,” the critics wrote. (Recent federal guidelines warn against both over- and underidentification of students of color.)

Morgan says he agrees that, in some cases, identifying too many students of color may be a legitimate concern. But he worries that districts are being encouraged to pay too much attention to raw differences among groups.

He pointed to the recent finding that schools in Texas had systematically denied students special education services in order to hit accountability targets.

“You quickly can enter a situation where schools are paying attention to those numerical targets and not identifying children who may have disabilities,” he said.