But Black students have lived with their racial identity. They know it can matter. They didn’t know that it would matter in this instance. But they knew it could. Thus their predicament. To be safe, not to be misled, they were less trusting. They stayed in churn.

Such caution has its costs. There could be value in that feedback. It might help them with their essay and their writing generally. Trusting less means they’d miss out on that benefit. And this would happen over and over in a diverse school.

In contrast, in Renick’s early years at GSU he knew of several low-income students who were dropped from school for not meeting only one small bureaucratic requirement, never to be heard from again. These kinds of experiences brought Renick to an important insight: Academic ability was rarely the reason low-income students dropped out of college. It was mostly little things, potholes along the road to graduation. As Andrew Grumbel put it in Won't Lose This Dream, his compelling book describing the transformation of GSU: “Contrary to conventional wisdom, academic performance was not the most important predictor of student success or failure.

A ten-member University Innovation Alliance has committed to exploring its data-driven strategies. Also, GSU has its own National Institute for Student Success, headed by Renick, that helps regional public universities use this approach. As the degree of implementation progresses in these schools, graduation rates have increased, sometimes drastically so, especially for low-income and minority students. And there is another sector of American higher education that, in taking a more developmental approach to its students, has long used its own versions of GSU-type supports: historically Black colleges and universities (HBCUs).

There is, of course, another pedagogical mindset in graduate education: Give students the freedom to find what they care about, a true passion that will sustain them as their careers unfold. In this mindset, it’s reasonable to rid programs of as much structure as possible to allow students’ own talents, ingenuity, and interests to guide their development. This approach has produced generations of distinguished scientists and scholars. But it may not work as well for a diverse student body. The ambiguities it allows into the training experience may impose a special burden on students whose identities put their belongingness under suspicion. The open-ended freedom assumed to foster development could leave them in near-constant churn.

It’s important to stress that the UC Berkley College of Chemistry, like GSU, used multiple practices to reduce ambiguity and give support in the setting. Multiple red buttons. Any one practice alone might not be enough. Take procedural justice— the degree to which a community adheres to nondiscriminatory, transparent practices in matter of hiring, promotion, and adjudicating disputes.

What he needs when he gets to school is something other than cognitive accountability— its good intentions notwithstanding. He needs signals that address his churn and that help resolve it: signs from adults and fellow students, for example, that he belongs in that school, that he is valued for who he is and who he can be, and that he is seen as having the potential to succeed there. And it wouldn’t hurt if he found some things about school interesting. For example, if his homeroom teacher read the class a compelling story every morning to get the day rolling. He needs signals that despite his status— and his churn to the contrary— he can trust the school’s beneficence toward him.

Yet in his school— very likely in the grip of cognitive accountability— he isn’t likely to have such experiences. Rather (as we shall see), he’s likely to get a more basic skills curriculum not much connected to his life, needs, or interests; a rigid disciplinary system; and frequent assessments of his cognitive skills that, from his perspective, may reinforce his image as a poor student— to himself and others.