The Mismatch

I want to be clear about something before I make my argument: I care about this stuff. I am brown. I am an immigrant. I have a biracial kid. I care whether our schools produce diverse leadership pipelines. I care whether kids from my community get a real shot. And I care whether the people who run our schools are actually thinking about outcomes, not just checking boxes and calling it progress.

Which is why I keep coming back to the mismatch research, and why it keeps me up at night.

There is a body of research that has been building for roughly two decades, and it says something that nobody in the diversity-and-access conversation wants to hear: putting underprepared students into selective educational environments does not help them. It harms them. And the people who suffer most are the ones the policy claimed to help.

The research is called mismatch theory, and its most documented version comes from Richard Sander's work at UCLA Law School. What the data consistently shows is this: students admitted to schools where they are significantly below the median academic preparation tend to perform worse than comparable students at schools where they fit the academic profile. They earn lower grades. They drop the hard courses — calculus, organic chemistry, advanced physics — that lead to STEM degrees. They lose confidence. And at the extreme, they drop out at higher rates than they would have if they'd attended schools where they were adequately prepared.

This is not a political argument. It is not a values argument. It is a measurement argument. And it is the argument that a body of research associated with Richard Sander at UCLA Law School has been making for roughly two decades — that mismatch is real, that it is consistent across law schools and at least some undergraduate contexts, and that the students who suffer from it are disproportionately the students admitted via preference policies designed to help them. I want to be honest about the limits of that claim, though. The mismatch hypothesis is contested in the published literature. Researchers like Zachary Bleemer at UC Berkeley and Princeton, and Richard Lempert at Michigan Law, have published rebuttals arguing the effect does not hold up in their data, and the question remains genuinely open in sociology, economics, and law. I'm persuaded by the mismatch side of the dispute. But "I'm persuaded" is the right level of confidence — not "the science is settled."

Here is what Philadelphia's own data shows, documented in a 51-page report by the Masterman Home and School Association published in February 2023. The authors — a multiracial group of Masterman parents — were not conservative think tanks. They were not anti-diversity activists. They were parents at the school watching what was happening to their children's education in real time. And what they documented was this: math proficiency dropped from 86% to 76% after the new admissions policy. Reading proficiency fell from 97% to 90%. Students with basic or below-basic competency levels were being admitted to a school that the report itself described as "not well-suited to accommodating a broader range of academic needs" at the high school level. At least one classroom teacher was reassigned to support struggling students. Language offerings for middle schoolers were cut. The pace of instruction slowed in some classes. And the report's conclusion, endorsed by a multiracial group of parents, was that the school was being "systematically dismantled".

I have a friend who teaches at a similar selective school in Philadelphia — one that went through the same lottery-based admissions change — and his account matches the report exactly. Students arriving who are not academically prepared for the pace and rigor of the curriculum. Teachers spending time on remediation that used to go to enrichment. The overall effect, as the HSA report put it, being to ask less of and deliver less to the students who were already the most academically prepared.

Now let me connect the two things that don't get connected in the public conversation about school diversity. The stated goal of the ZIP-code preference policy was to increase representation of Black and Hispanic students at schools like Masterman. I share that goal. I want diverse leadership pipelines. I want kids from every ZIP code to have a real pathway to the top. But here is the problem: the students being admitted under the new system are not the students who would have been admitted under a real socioeconomic preference program. They are students who meet a minimum threshold from specific ZIP codes, which means they are students who may be significantly less academically prepared than their peers. And when you put a student who is at the 80th percentile of academic preparation into a school where every other student is at the 99th percentile, you do not help that student. You put them in an environment where they will struggle, where they will be reminded every day that they do not belong at the same table as the students around them, and where the path of least resistance is to avoid the hardest classes rather than to take them.

The mismatch literature shows this explicitly in STEM. When Black students at selective universities are admitted with significantly lower preparation than their peers, they are more likely to switch out of STEM majors — not because they lack ability, but because the signal they receive from being surrounded by better-prepared peers tells them they don't belong. The result, over time, is fewer Black doctors, fewer Black engineers, fewer Black physicists. The policy designed to increase representation in elite spaces ends up decreasing it, because the students who get in under the new system don't survive the environment they're placed in.

And there is a second, subtler problem that almost nobody talks about. When you admit students at the 80th percentile alongside students at the 99th percentile, you create academic stratification within the school. The students at the top end and the students at the bottom end are not the same. They have different preparation levels, different study habits, different comfort with the material. And social groups, in high school and college, tend to form around academic similarity. This is well-documented: students who are at similar academic levels spend more time together, form tighter bonds, end up in the same classes and the same study groups. Before the lottery system, everyone at Masterman had made the same cut — they were all at the 99th percentile, and their peer groups reflected that diversity. Now, with students entering at different preparation levels, the social dynamics are changing. The friend groups that used to cut across race — because everyone was at the top tier — are now starting to stratify by academic tier. And academic tier, in a city with Philadelphia's history of residential segregation, will correlate with race.

I think about this differently than a lot of people in my political camp, maybe. I do not believe that good intentions produce good outcomes. I believe in measurement. I believe in evidence. And the evidence says that what Philadelphia is doing with its selective school admissions is not going to produce the diverse pipelines of doctors, engineers, and leaders that the people advocating for this policy say they want. It is going to produce something else: kids who arrived unprepared being sorted into academic tiers where they don't belong, losing confidence in the process, and exiting the pathways that would have led them somewhere real.

The people who designed this policy had a goal, and I think the goal was sincere: more diverse representation at the city's best schools, more pathways for students from neighborhoods that had been historically excluded. But they did not do the research, or they ignored it. The mismatch literature goes back roughly two decades. The Duke University research on STEM persistence and underprepared students is not obscure. The evidence that putting students in environments where they're significantly underprepared harms them rather than helps them has been replicated across contexts. And Philadelphia's own data from Masterman — in a report written by the school's own parents, not by ideological opponents — confirms it.

If you want to help students from underrepresented neighborhoods, the answer is not to lower the bar at selective schools. It is to fix the schools those students come from — the AC-less buildings, the overworked teachers, the absent counseling infrastructure — so that when those students arrive at a school like Masterman, they are genuinely prepared for the work. Anything else is theater. And the people who pay the price for the theater are the students who were told they were being helped, and who discover too late that they were being set up to fail.