Black students in Louisiana are suspended for slightly longer than white students after being involved in the same fight, according to new research that adds to a roiling national debate about school discipline.

The study comes as the U.S. Department of Education appears to be considering rescinding Obama-era guidance on school discipline. Its findings — that black students are treated more harshly — bolster the case of civil rights groups that want the guidelines to remain, noting that suspension rates for America’s black and poor students remain disproportionately high.

“Given that we find that direct discrimination occurs in this context, with a black and white student receiving different punishments for the same exact incident, it seems likely that direct discrimination would [also] occur where discipline disparities are less visible,” the researchers write.

Still, the difference between suspensions given to black and white students was quite small, amounting to a fraction of a school day.

The research, released by the Education Research Alliance at Tulane University, used a trove of discipline data from Louisiana that stretched from 2001 to 2014. Consistent with past research, the study finds that black and poor students have substantially higher suspension rates than white and more affluent peers.

Figuring out why is tricky. Is it because certain groups of students behave differently, or because teachers and administrators respond differently to the same behavior?

The study can’t rule out the possibility that the small difference in length of suspensions was due to factors like whether one student instigated the fight. But the researchers argue that a fight between two students is an objective offense where school officials would be expected to treat participants similarly. The fact that even in this context there is evidence of bias, however small, is worrisome, they say.

The study is directly relevant to the debate about the 2014 federal guidance, which warned schools that the Department of Education might initiate civil rights investigations “based on public reports of racial disparities in student discipline combined with other information.” U.S. Education Secretary Betsy DeVos recently said she was “looking closely at” this guidance, and conservatives have urged her to scrap it altogether on the grounds that it has contributed to disorder in schools.

Much of the debate turns on how to interpret well-documented disparities in suspension rates.

Mike Petrilli, a critic of federal efforts to monitor school discipline and the head of the Fordham Institute, a conservative think tank, analogizes the debate to the discussion around the test-score “achievement gap.”

“Nobody would say that we’re going to look at these achievement gaps and say that all of the achievement gap is because teachers are racist,” he said. “We might say part of that is.”

Petrilli suggests that the guidance has schools fearing that they risk being accused of discrimination based just on suspension differences across student groups.

Catherine Lhamon, who served as the assistant secretary for civil rights at the education department under Obama and is now the chair of the U.S. Commission on Civil Rights, said that guidance doesn’t suggest that suspension rates should be interpreted as proof of discrimination. “Data alone is not evidence of discrimination,” she said, but it “is relevant to an inquiry.”

The Tulane study marks an attempt to see if discrimination really is at play. The researchers compared the length of suspensions given to black and white students for the same infraction in Louisiana. Black students, they found, are suspended for longer — 2.6 school days, compared to 2.2 days. This gap shrinks but still exists when comparing black and white students in the same school, grade, and school year.

That suggests, but does not prove, that there is discrimination involved in those decisions.

The researchers then examined suspensions after a fight between a black and a white student. They found that, even controlling for past incidents of fighting, black students receive longer suspensions — though the difference was very small, equivalent to black students facing one additional day of suspension for every 20 fights.

Past studies have come to conflicting results on the question of bias: Consistent with the theory of bias, black students are more likely to be suspended for those discretionary offenses than objective behaviors. Black teachers are also less likely to suspend black students compared to white teachers and perceive behavior of black students differently.

On the other side, there is evidence that the race-based gap in discipline is accounted for by past behavioral issues (though if those prior behaviors were reported with bias, that would invalidate these results).

The Tulane study points out that even if discipline disparities are not caused by direct bias at the school level, they likely are due to broader societal issues, including long-standing discrimination.

Even if schools are responding to genuine differences in behavior across groups of students, the authors write, “this exonerates neither schools nor broader societal forces from contributing to varying levels of misbehavior.”

As DeVos reconsiders, the debate intensifies

Meanwhile, Petrilli says he is trying to reduce the intensity of the debate, writing a blog post titled, “In search of common ground on school discipline reform.”

The debate seems to be growing more polarized instead.

Just recently, Petrilli organized a visit for teachers who said school discipline reforms had led to chaos in the classroom to speak to leaders at the U.S. Department of Education. Katherine Kersten, who wrote an essay critical of discipline reform efforts titled “No Thug Left Behind,” helped connect Petrilli to Minnesota teachers for the visit.

(Kersten said she didn’t write the title, though didn’t see a problem with it. “Some children are thugs,” she said in an interview.)

And an op-ed in the New York Daily News by Max Eden and Robert Pondiscio linked school discipline reform efforts in New York City to the recent stabbing death of a high school student. (There’s no evidence that changes in discipline policies led to this fatality, though they point to surveys showing a decline in students who felt safe at the school.)

Progressives are also on high alert about the Trump administration’s record on civil rights.

In an op-ed earlier this year, former Secretary of Education Arne Duncan linked student suspensions to a broader critique of the administration, writing, “Leaving enforcement of civil rights laws to states will breed chaos, undermine the education of millions of children, and subject students of every age to abuse, neglect, indifference and outright racism, sexism, and anti-immigrant hostility.”

While conservative critics of the discipline reform efforts have embraced teachers who say the policies have been poorly implemented, national teachers unions say they support the existing guidance.

“Black students are nearly four times more likely to be suspended than their white peers for the same infractions,” said Randi Weingarten, the president of the American Federation of Teachers, in a statement. “We must confront the racial bias that underlies that reality, which the current guidance attempts to do.” Weingarten has written about her own change of heart about harsh discipline policies.

Harry Lawson, the director of human and civil rights at the National Education Association, also said the union was opposed to rescinding guidance, but said some places that have implemented discipline reforms have hit challenges.

Even supporters have acknowledged as much. A report by the Advancement Project, a civil rights group that backs reduction in exclusionary discipline, for instance, found that the Miami-Dade school district’s promise to eliminate student suspensions led to the creation of “student success centers” that became “little more than warehouses for students who have been removed from school.”

Lawson said the key is training and resources to implement alternatives to suspensions. He said the NEA sometimes hears feedback from members akin to the teachers from Minnesota who spoke to Department of Education officials — but that they weren’t representative.

“What we also hear more of is a recognition from our members that there is something wrong [with exclusionary discipline],” he said. “And generally what they are asking is, what do we do … and can you all support us?”