I was a doctoral student in Denver, in the back of a school board meeting, when I watched twenty-five years of my life get decided in about nine seconds. A group of community representatives had finished presenting. They had laid out the findings of the scholars I had spent years reading, the people whose work I could feel in my chest. A board member leaned forward and asked where the data was. A presenter said, this is the data. The board member said, where are the numbers.
Then nothing.
Silence.
I have carried that silence for over twenty-five years. This essay comes out of it.
The room had just been handed the lived and documented knowledge of whole communities, gathered by the people who had lived it, and it did not count. Not because it was wrong. Because it did not arrive in numbers. The evidence was not missing. It was in the wrong language for that room.
For seven episodes this season has traced a practice older than its own name. Last time I told you about the scholars who proved that practice in the language of story, past any honest doubt, at kitchen tables and in homes and in schools the country had already closed. I ended by saying there was one thing they had not yet done. They had not put a number on it that a legislature could not wave away. This is the episode where that happens. I have to tell it in the first person, because I am one of the people it happened to.
The question I could not shake
A few weeks after that board meeting, a classmate gave a presentation citing Christopher Swanson, then at the Urban Institute. Swanson had looked at the graduating class of 2001 and concluded that the country had been overstating how many students actually finish high school. So he built his own measure from the federal data and followed students grade to grade, asking how likely a ninth grader was to walk out four years later with a regular diploma. No GED. No other route. Nationally, about sixty-eight percent. For white students, about seventy-five. For Black students, fifty. For Native students, fifty-one. For Mexican and other Latino students, fifty-three. For Asian students, seventy-seven.
Sit with the Black number. Half.
I want to be careful here, because the season’s whole argument lives in this spot. Swanson was not a careless researcher, and he did not set out to harm anyone. What I can tell you is what happened when I, a Black man from a Black neighborhood in Texas, heard that number. I could not make it fit. Where I come from, finishing high school is an occasion. Family drives in. People take off work. So I did something that was not a research method. It was a man trying to line a number up against his life. I started counting, asking everyone I knew the same question. How many people do you know who did not finish high school. The highest answer anybody gave was around thirty percent.
That was when I understood that I was still standing in the shadow of the Coleman Report. Not the study. What it taught the country to expect. A field so used to cataloguing what was broken in us that a claim that half of us failed could be repeated by educated people, and no one stopped to ask whether it was true.
And it was not only the Black number. Those were the years of the Black and white gap and the Mexican and white gap. But look at Swanson’s own table. Asian students finished at seventy-seven, white students at seventy-five. A white and Asian gap sat right there, and almost no one named it or built a research industry around fixing white students. The word gap is not a measurement. It is a decision about who needs explaining.
So I made a choice. I took every statistics course I could find. Not because I had stopped believing the stories. Because I meant to carry a cultural context into a place that had never held one.
A cultural lens, applied to a number
My dissertation went into a national dataset and to the transcripts, the actual record of what students finished. Same children, same years, a different answer. In my analysis, Black students finished at fifty-nine percent and white students at seventy-two. Nine points higher for Black students than the number in that classroom.
A number is not the end of an argument. It is the sum of the choices that built it.
So let me show you the choices. Start with what Swanson chose not to count. No GED. No alternative route. In communities where a young person may work a job, help raise a sibling, and finish on a different road, a rule that counts only the four-year diploma does not measure those students. It erases how they actually arrive. Then look at which states were in the data at all. To calculate a rate the way he did, a state had to report four straight years to the federal collection, and several of the largest centers of Black life did not. A study about Black graduation, built on a map with much of Black life left off it.
That is not a conspiracy. It is quieter and more ordinary. When the people who design a measure do not live inside the community the measure is about, the holes in it are invisible to them and obvious to us. That is the whole argument of this show, standing in a spreadsheet.
The variable finally gets a name
While the graduation argument ran through the early 2000s, something else landed. In 2004 an economist named Thomas Dee used a randomized experiment, the Tennessee class-size study, where students and teachers had been assigned to classrooms by chance. That let him ask a question the field had never answered cleanly. What happens when a student has a teacher of their own race. His answer: Black students taught by Black teachers scored higher in both math and reading.
Read that against everything in this season. The variable nobody wrote down for a century. The column that kept coming up missing. In the most rigorous design the field had, it finally showed up. And not many people were listening.
I made my own choice then, and it came straight from the scholars of the last episode. My research would center what worked for students of color, not what did not. Foster, Ladson-Billings, Moll, González, and Valenzuela had already taught me where to point the lens. I could not put one question down. Back home in Texas, I kept meeting people from my own blocks who had earned doctorates. Nothing in the deficit literature explained my street. One thing did. The variable.
Dee’s window was small, mostly the earliest grades. I had trained on large national datasets, on longitudinal modeling and latent class analysis, so I leaned into the data to ask whether what he found held up past kindergarten. With Chance Lewis and Yubo Zhang, in 2009, I published a study of ethnic matching and reading. Two years later, with Colleen Eddy, I followed matching and mathematics from kindergarten through fifth grade.
What it cost to ask
Now the part that does not usually make it into the citation.
It was hard to find a home for the work. From some qualitative scholars, the response was, essentially, tell us something we did not already know. They were not wrong that the communities knew. That is this whole season. From some quantitative scholars, the objection was stranger. One told me the work was not serious because I had cited more qualitative researchers than quantitative ones. Sit with that. The evidence of a hundred years, gathered by the people who lived it, counted against me.
Then it came from inside my own building. In an annual review, my department chair and my dean told me that if I kept on this line of research, I would not be taken seriously. I was untenured, with a family and a mortgage. On one side, I knew the work was right, in the data and in the culture, which to me is the higher bar, not the lower one. On the other side was the plain arithmetic of a career. That is the tax this season is named for, arriving in my own life, in a performance review.
What changed it was a person. Richard Milner came in as editor of Urban Education, looked at the work, saw it, and told me to keep going. Not a new method. Not a better dataset. One scholar with the standing to open a door, deciding the question was worth the space. I am here because of one.
Where is the data. Here is the data.
Let me answer that board member, twenty-five years later.
For Black students, the finding is now among the steadiest in all of education research. Dee opened the door in 2004. The reading study came in 2009, the mathematics study in 2011, and what we found compounded: the more years a Black child spends with a Black teacher, the larger the advantage in their path. Then the rest of the field arrived. Egalite, Kisida, and Winters in 2015. A full review in 2019 found the pattern holding across study after study. And the longest view of all found that a Black child who has a Black teacher in the early grades is more likely to finish high school and more likely to enroll in college. A relationship in first grade, visible a decade later in whether a young person walks onto a campus.
For Mexican and other Latino students, the evidence has moved the same way, in reading and in mathematics, and past the classroom teacher to the staff across a whole building. It even holds for children learning in two languages, where you might expect the gain to be about language alone. It is not. When the teacher shares the child’s bilingual world, the child gains in mathematics too, not only in words. The child is not being translated. The child is being known.
For Asian students, everything depends on refusing to treat a whole continent as one category. In the Seattle area alone there are more than sixteen distinct Asian communities. Look at tenth-grade mathematics in Seattle in 2007 and 2008. Chinese, Japanese, and Korean students scored well above the Asian average. Cambodian, Hmong, Mien, and Laotian students scored nearly eighteen percent below it. One category, eighteen points of difference living inside it. Report only the average and you erase both ends, and you miss the children who most need to be seen. Now put matching on top of that picture. The 2015 Florida study followed every test-taking student, grades three through ten, across eight years. And the single largest match effect anywhere in it, larger than for any other group, was for Asian students in mathematics, in the very group the country had already decided did not need help. It fits what the Seattle data would predict. The more unfamiliar the institution, the more a matched teacher can carry the child. That is what matching is.
For Native students, I have to be honest about the evidence, because honesty is the point of this episode. The base is smaller, partly because the population is smaller, and partly because lumping hundreds of distinct tribal nations into one category repeats the exact error we just named with Asian students. What the research does show is steady. Native students who have teachers who understand and respect their community context show stronger engagement and stronger performance. And for these students a matched teacher carries something more, because their communities faced a federal program built to erase their languages and their nations. A teacher who supports cultural continuity is not offering an accommodation. That teacher is answering a specific historical harm.
The newest work has moved past test scores entirely, which is overdue. A matched teacher changes who gets disciplined, who shows up to class, who gets referred for services, and who gets recommended for gifted programs. Show up, be seen, get referred. None of that is a test score, and all of it is a life.
I owe you the limits too, because a finding you tell only one way is a slogan. Researchers have named the diversity tax, the extra cultural and emotional labor loaded onto teachers of color who become the counselor, the family liaison, and the translator for a whole school on top of a full teaching load. There is a real risk of essentialism, of treating a teacher’s identity as a stand-in for their skill. There is the danger of hiring teachers of color, giving them no support, and then watching them leave. The match is not a substitute for building a school worth staying in.
Where this leaves us in 2025
Which brings me to now, and I will not dress it up. In the last few years the equity infrastructure that districts spent years building has come under an organized political attack. Laws restricting how race can be discussed in classrooms. Offices closed. Diversity hiring goals challenged, and in places prohibited. Grow Your Own teacher programs losing the budget lines that funded them.
I want to name the argument, because I have heard it before. The claim that equity work is ideology rather than a response to documented inequity is not new. It is the same argument made against desegregation, and against multicultural education. The research in this episode makes that argument wrong. Being wrong has never stopped an argument from winning, when the people who translate research into policy prefer a different conclusion.
So here is where I have landed after twenty-five years. The board member’s question was fair. Where is the data is a fair thing to ask. What was not fair was that the answer he received, the lived and documented knowledge of entire communities, did not count until people like me spent a decade rebuilding it in his language. The numbers were never the problem. The problem was who got to decide what counted as a number, and whose life had to be translated before it could be believed.
Where we are going
We have the data now. The variable that went unwritten for a century has a name, a method, and a shelf of studies behind it. We answered the board member. Where is the data. Here it is.
But having the data is not the same as finishing the work. The country still fills its classrooms with children of color and staffs them with a teaching force that does not look like them. That match, the one we can now measure so precisely, is still missing from the school day for most of these children.
That is why the newest chapter of my own work, with Jemimah Young and Devin Williams, is titled Still Searching for a Match. Still. After everything in this episode. The question is no longer whether the match works. We have settled that. The question is why, knowing what we know, we still have not built the hiring, the schools, and the routes that would put a matched teacher in front of every child who needs one.
The data was never the hard part. The will is.
Next time, we take this out of the journals and into the buildings, and we ask the harder question. We know what the match does. So what would it actually take to give it to every child who needs one, and why, after all of this, have we still not done it. I’ll see you there.
DEB
