Michelle Rhee
Former D.C. schools chancellor Michelle Rhee. Credit: Commonwealth Club/Flickr

The following is in response to “Hot for Teachers,” from the June/July/August 2017 issue.

To remain aloft, a hot air balloon must be fed regular bursts of hot air. Without hot air, the balloon falls to earth. That seems to be the appropriate analogy for the District of Columbia Public Schools (DCPS) during the ten-year regime (2007–2016) of Chancellors Michelle Rhee and Kaya Henderson. Their top-down approach to school reform might not have lasted but for the unstinting praise provided by influential supporters from the center left and right—their hot air. The list includes the editorial page of the Washington Post, former U.S. Secretary of Education Arne Duncan, and philanthropist Katherine Bradley. The most recent dose is “Hot for Teachers,” in which Thomas Toch argues that Rhee and Henderson revolutionized the teaching profession in D.C. schools, to the benefit of students. But this cheerleading obscures a harsh truth: on most relevant measures, Washington’s public schools have either regressed or made minimal progress under their leadership. Schools in upper-middle-class neighborhoods seem to be thriving, but outcomes for low-income minority students—the great majority of enrollment—are pitifully low.

Toch is an engaging storyteller, but he exaggerates the importance of positive developments and misrepresents or ignores key negative ones, including dismal academic performance; a swollen central office bureaucracy devoted to monitoring teachers; an exodus of teachers, including midyear resignations; a revolving door for school principals; sluggish enrollment growth; misleading graduation statistics; and widespread cheating by adults.


When they arrived in 2007, Rhee and her then deputy Henderson promised that test scores would go up and that the huge achievement gaps between minority and white students would go down. Here’s how Toch reported what has happened on their watch: “While Washington’s test scores have traditionally been among the lowest in the nation, the percentage of fourth graders achieving math proficiency has more than doubled on the National Assessment of Educational Progress (NAEP) over the past decade, as have the percentages of eighth graders proficient in math and fourth graders proficient in reading.”

Those results, however, stop looking so good once we disaggregate data about different groups of students. Despite small overall increases, minority and low-income scores lag far behind the NAEP’s big-city average, and the already huge achievement gaps have actually widened. From 2007 to 2015, the NAEP reading scores of low-income eighth graders increased just 1 point, from 232 to 233, while scores of non-low-income students (called “others” in NAEP-speak) climbed 31 points, from 250 to 281. Over that same time period, the percentage of low-income students scoring at the “proficient” level remained at an embarrassingly low 8 percent, while proficiency among “others” climbed from 22 percent to 53 percent. An analysis of the data by race between 2007 and 2015 is also discouraging: black proficiency increased 3 points, from 8 percent to 11 percent, while Hispanic proficiency actually declined, from 18 percent to 17 percent. In 2007 the white student population was not large enough to be reported, but in 2015 white proficiency was at 75 percent.

The results in fourth grade are also depressing. Low-income students made small gains, while “others” jumped to respectable levels. As a consequence, the fourth-grade proficiency gap between low-income and “other” students has actually increased, from 26 to 62 percentage points, under the Rhee/Henderson reforms.

Results of the Common Core tests known as PARCC, first administered in 2015, are similarly unimpressive. The black/white achievement gap is 59 percentage points. Although DCPS students achieved 25 percent proficiency system-wide, the average proficiency in the forty lowest-performing schools was 7 percent. In ten of the District’s twelve nonselective, open-enrollment high schools, somewhere between zero and four students—individuals, not percentages—performed at the “college and career ready” level in math; only a few more achieved that level in English. This is a catastrophic failure, strong evidence that something is seriously wrong in Washington’s schools.

Remember that these students have spent virtually their entire school lives in a system controlled by Rhee and Henderson. In short, despite promises to the contrary, the achievement gap between well-to-do kids and poor kids as measured by the NAEP has widened under their watch and is now over twice as high in fourth grade and two and a half times as high in eighth as it was a decade ago. White proficiency rates now run 55 to 66 percentage points above black proficiency rates and 42 to 66 percentage points above Hispanic rates.

Toch asserts, “Scores have risen even after accounting for an influx of wealthier students.” However, evidence suggests that the test score increases in some grades are most likely a by-product of gentrification. The percentage of white test takers has increased steadily over the last decade (5 percent to 16 percent in fourth grade and 5 percent to 9 or 10 percent in eighth grade), as has the percentage of Hispanic students (9 percent to 16 percent in fourth grade and 7 percent to 15 percent in eighth grade). In the nation’s capital, almost all white kids are from well-to-do families, while Hispanic and black kids in public schools are mostly from low-income households. A 2015 report by the National Research Council pointed out that most of the recent academic gain was likely the result of more affluent families moving into Washington and enrolling their children in public schools.

Central Office Bloat 

Under Rhee and Henderson, spending on non-teaching personnel has swollen dramatically. According to the latest statistics from Census Bureau fiscal reports, DCPS central office spending in 2015 was 9.5 percent of total current expenditures, compared to 1 percent or less in surrounding districts. Today DCPS central offices have one employee for every sixty-four students, a striking change over the pre-Rhee/Henderson-era ratio of 1 to 113 students. Those central office dollars could have been used to provide wraparound social services for children, services that would have allowed teachers to be more effective.

Many of these highly paid non-teachers spend their days watching over teachers in scheduled and unscheduled classroom observations, generally lasting about thirty minutes—not even an entire class meeting. Why so many of these teacher watchers? Because those who subscribe to top-down management do not trust teachers—which leads to the next part of the DCPS story that its admirers misrepresent, downplay, or ignore.

Teacher Turnover and Resignation

Toch writes about Washington’s success in recruiting teachers, even poaching them from surrounding districts. He attributes this to higher salaries and increased professional respect and support. And he adds, in a carefully qualified sentence, that “the school system’s strongest teachers are no longer leaving in droves for charter schools.” Well, perhaps they’re not leaving for charter schools, but they sure as heck are leaving—in droves. Toch fails to mention the embarrassingly high annual turnover of 20 percent system-wide and a staggering 33 percent every year over the last five years in the forty lowest-performing schools. This means that in the neediest schools, one out of every three teachers is brand new every year. And all newly hired teachers, whether novices or poached from elsewhere, leave DCPS at the rate of 25 percent annually. In a recent study of sixteen comparable urban districts, the average turnover rate was just 13 percent.

Defenders of the D.C. approach would have you believe that these teachers have failed to increase test scores. While that is true in some cases, other evidence should be considered. Student journalists at Woodrow Wilson High School interviewed this year’s departing teachers, who expressed frustration with “DCPS’s focus on data-driven education reforms” and “lack of respect and appreciation.” Teachers, including those rated “highly effective,” cited the stress of frequent changes in the demands of the IMPACT teacher evaluation system as well as the absence of useful feedback.

Nor does Toch discuss the large number of resignations during the school year, always a sign that something is amiss. The Washington Post reported recently that nearly 200 teachers quit their jobs after the school year began. “It is an emergency when a quarter or more of the teachers in some schools have resigned during the school year,” D.C. council member Robert C. White wrote in a letter to the chairman of the council’s education committee. According to the Post’s analysis, half the schools in the system had three or more midyear resignations.

The Revolving Door of the Principal’s Office

Every year about 25 percent of DCPS schools open with a new principal. Research finds that principal turnover generally results in lower teacher retention and lower student achievement, particularly at high-poverty and low-achieving schools. Research also suggests that principals must be in place for at least five years to accomplish large-scale change, but only twenty-five of the 115 DCPS schools in the 2016–17 school year had principals who had served five years or more, and most DCPS schools have had two or three principals in the past five years. Unfortunately, the greatest upheavals are in schools serving large numbers of low-income children, kids who need stability wherever they can find it.


Toch makes much of the recent enrollment increase, calling it evidence of the success of the Rhee/Henderson approach. It’s true that PS enrollment, after bottoming out at 44,718 in 2009, had risen to 48,555 in fall 2016. But over the same span, charter enrollment grew by more than 11,000 students. If parents are voting with their feet, then DCPS acolytes ought to be concerned.

Deceptive Graduation Numbers

Toch cites DCPS graduation rates as having climbed to 69 percent, “the highest in the city’s history.” But even education reform enthusiasts dispute the validity of claims about graduation rates. Analyst Robert Pondiscio calls graduation percentage “the phoniest statistic in education,” and veteran Washington Post reporter Jay Mathews suggests that reported improvements are “a mirage.” They argue that increased pressure on teachers and principals to pass students, and highly suspect “credit recovery” practices that allow students to complete a semester’s course work online in a week, combine to inflate the graduation numbers. By some estimates, at least 25 percent—and perhaps as many as 50 percent—of DCPS’s graduating seniors are helped across the line by at least one credit recovery class.

Even if the graduation rate increase is real, given the minuscule percentages of students in most high schools who are proficient by the Common Core PARCC standards, what significance could it possibly have? If undereducated students are being graduated, how is that anything but a hollow triumph?

Cheating by Adults

Toch dismisses D.C.’s cheating scandals in just two sentences: “In March 2011, USA Today ran a front-page story headlined ‘When Standardized Test Scores Soared in D.C., Were the Gains Real?,’ an examination of suspected Rhee-era cheating. The problem turned out to be concentrated in a few schools, and investigations found no evidence of widespread cheating.”

Contrary to Toch’s assertions, cheating—in the form of suspiciously high rates of erasures of wrong answers and filling in the right ones—occurred in more than half of DCPS schools. The changes were never thoroughly investigated beyond an initial analysis by the agency that had corrected the exams in the first place, CTB/McGraw-Hill. Deep erasure analysis was never ordered by Rhee, her then deputy Henderson, or the mayor. The “investigations” Toch refers to were either controlled by Rhee and, later, Henderson or conducted by inept investigators—and sometimes both. The first two were conducted by a company that had failed to detect cheating in Atlanta, the epicenter of cheating. Rhee, and subsequently Henderson, tightly controlled the inquiries, limiting the number of schools that could be visited, the number of interviews that could be conducted, and even the questions that could be asked.

Why would so many schools be driven to cheat? In her one-on-one meetings with all her principals, Rhee insisted that they guarantee test score increases and made it clear that failing to “make the numbers” would have consequences. The adults who subsequently changed answers, coached students during testing, and shared exams before the tests were intent on keeping their jobs, which depended on higher test scores. (More details can be found at themerrowreport.com, in a blog post entitled “‘School Reform’s Hot Air Balloon.”)

What Has Gone Well

The D.C. public school system has improved in a number of ways since Rhee became chancellor in 2007, and both Rhee and Henderson deserve recognition. Full-day preschool for three- and four-year-olds has been greatly expanded. Facilities have had extensive, badly needed repairs, and a majority of schools have undergone complete modernizations. Pupil/staff ratios are smaller, more schools have extended-day or extended-year programs, and schools offer more electives, activities, and sports. More students participate in AP courses and dual-language programs than did ten years ago. Teacher salaries remain significantly higher than in almost everywhere else in the country.

But, ultimately, Rhee and Henderson lived and died by test scores, and their approach—more money for winners, dismissal for losers, and intense policing of teachers—is wrongheaded and outdated. Their conception of schooling is little changed from an industrial age factory model in which teachers are the workers and capable students (as determined by standardized test scores) are the products. The schools of the twenty-first century must operate on different principles: students are the workers, and their work product is knowledge. This approach seeks to know about each child not “How smart are you?” but, rather, “How are you smart?”

Rhee and Henderson had the kind of control other school superintendents can only dream of: no school board, a supportive mayor, generous funding from government and foundations, a weakened union, and strong public support. Yet, despite carte blanche to do as they pleased, they failed. Without the hot air of public praise, the Rhee-Henderson balloon would have plummeted to earth.

Toch’s article indicates that the stream of hot air will continue to keep the DCPS balloon aloft. The stakes are high, because other like-minded reformers around the country hold up DCPS as “proof” that a top-down, test-centric approach is the key to improving public education. It appears that under Henderson’s successor, Antwan Wilson, the failing Rhee-Henderson approach will continue. That’s sad, because Washington’s students and teachers deserve better.

Thomas Toch responds:

Near the top of my story on the District of Columbia Public Schools’ work over the past decade to improve teachers and teaching in the nation’s capital, I pointed out that the conventional characterization of the DCPS effort is badly outdated. I wrote that when most people think of D.C. school reform, they focus on former Chancellor Michelle Rhee’s early efforts to fire bad teachers, and ignore everything that happened after Rhee left the school system in 2010.

Mary Levy and John Merrow, longtime Rhee opponents whose arguments align closely with those of the nation’s teachers’ unions, have made my point for me. They rehash old grievances and ignore key reforms that Rhee’s successors have implemented in the seven years since her departure, reforms that are the focus of my story.

I found in the course of my research that Rhee’s successors went well beyond her early efforts to remove weak teachers. While teaching in the United States has long been a low-status occupation, Chancellor Kaya Henderson and her staff effectively transformed teaching in Washington into a performance-based profession that provides recognition, responsibility, collegiality, support, and significant compensation—features that policy experts and Rhee critics have long sought.

They introduced new recruitment, retention, and staffing policies that have improved the flow of talented teachers into the city’s classrooms, provided teachers with a wide range of professional opportunities that are rare in public education, developed a highly regarded new curriculum and trained teachers extensively in how to deliver it, and raised compensation dramatically, leaving charter schools struggling to compete.

The problem that Levy and Merrow seem to have with these reforms is that they include policies that teachers’ unions and others on the left don’t like, including comprehensive teacher evaluations, the abandonment of seniority-based staffing, and performance-based promotions and compensation.

There’s little coherence to their critique. Why do they refight an old battle about cheating on standardized tests that took place in the Rhee era? Curious readers can consult a 2015 National Research Council report, which concluded that “from what we could determine . . . the alleged [cheating] violations were likely not widespread enough to have affected citywide scoring levels.” Henderson and the mayor’s office toughened testing security shortly after the problem surfaced. Besides, the extent of cheating seven years ago is tangential to my piece, which was about the transformation of the teaching profession in D.C. I mentioned the controversy only as an example of how DCPS critics fixate on the Rhee era and ignore what has happened since—as Levy and Merrow proceeded to do.

Then they veer in another direction, complaining about excessive central office spending—“bloat,” they call it. But they neglect to tell their readers that before Rhee, the patronage-plagued DCPS central office couldn’t even calculate daily attendance, much less educate students. There were no teaching standards or common curriculum. Many teaching vacancies weren’t filled until September or October, and new teachers often didn’t get paid for months. New textbooks gathered dust in warehouses while there weren’t enough to go around in classrooms. Elementary schools mostly didn’t teach art or music. High school electives were rare. And the system was hemorrhaging students and top teachers to charter schools.

Rhee and Henderson brought in a senior team with talent rarely matched in public education. They used substantial federal and foundation funding to build needed infrastructure and to support reform. They created DCPS’s first-ever district-wide teaching standards, developed the most comprehensive teacher evaluation systems in the nation, built a sophisticated career ladder that promotes teacher leadership roles, and designed a performance-based pay system that has lifted top salaries from $87,000 to $132,000 (and even higher in year-round schools).

They established rigorous new academic standards, drafted a comprehensive new curriculum built on those standards, created hundreds of model lessons to help teachers deliver the new curriculum (including dozens of the city’s top teachers in the design work), revamped the district’s principal and teacher recruitment and hiring systems, replaced the district’s information-management platforms, rebuilt DCPS’s long-dysfunctional special-education system, and, last summer, introduced an entirely new school-based teacher training system and a curriculum to support it, giving teachers the kind of collaboration and support that many have long said they want but have rarely received.

Levy and Merrow ignore this work. Instead, they claim that much of the central office budget is spent on “teacher watchers,” subject-matter experts who observed teachers under the comprehensive new teacher evaluation system. They contend that evaluating and coaching teachers is tantamount to not trusting them, which is a bit like saying a baseball team that hires a coach must not trust its players. What’s more, Henderson and her team discontinued the use of the school system evaluators a year ago in favor of a new school-based teacher training and support program (though many teachers valued being rated by independent experts rather than merely by their principals).

Levy and Merrow claim that the city has a 25 percent teacher “turnover rate” and that it’s evidence of the failure of Henderson’s strategies. What they don’t say is that roughly half the “turnover” is teachers moving to other DCPS schools, often for leadership opportunities. Another 2 percent are voluntary departures among low-rated teachers. And another 3 percent are teachers fired for poor performance—something that virtually never happened in D.C., or anywhere else in public education, in the pre-Rhee era. In contrast, among the 37 percent of DCPS teachers earning the city’s top ratings, annual attrition is a mere 6 percent, lower than the turnover rates of many of Washington’s elite private schools.

Levy and Merrow also point to a recent Washington Post article as evidence of high midyear teacher attrition. The paper reported that 184 of the school system’s 4,150 teachers departed during the 2016–17 school year. It turns out that 30 percent of that attrition was in three struggling schools and a fourth that runs year-round. The rest of the city’s schools averaged a little over one midyear departure apiece—hardly a big number given the demands of educating DCPS’s many deeply impoverished students, and a rate that is no higher than that of some elite Washington-area private schools. As the reporter acknowledged in the story, “In most DCPS schools, the faculty is stable,” and “the number who quit abruptly is small compared with the total [DCPS] workforce.”

Meanwhile, Levy and Merrow are silent about the fact that three times as many DCPS teacher recruits today are under contract by the end of the previous school year than in the pre-Rhee era (an important improvement, since studies show that early hires yield stronger teachers), that more new hires have previous teaching experience, and that DCPS’s recruitment reforms have resulted in higher-quality replacements for low-rated teachers, replacements who produce upward of five months’ worth of additional learning in math over three school years.

Levy and Merrow cite the district’s principal turnover rate as evidence of the failure of reforms. But that’s partly by design: in DCPS, unlike in most districts, principals can actually lose their jobs for underperformance. Unfortunately, principals also leave because in the toughest parts of Washington the work is extraordinarily demanding. As I mentioned in my article, Eric Bethel, the current principal of Turner Elementary, told me, “If you’re not driven by a sense of social justice, you’re not going to last.”

Then Levy and Merrow pull out a real oddity: Washington’s student enrollment trends. They say that I “make much of” DCPS’s rising enrollment in recent years as “evidence of the success of the Rhee/Henderson approach.” In fact, while I discuss enrollment growth at Turner, I don’t mention system-wide trends in my piece at all. If I had, I would have noted that enrollment has risen five years in a row, after four decades of decline. Levy and Merrow say that Henderson’s reforms are a bust because enrollment has grown less in DCPS than in charter schools. But five years of growth after forty years of hemorrhaging students is inarguably a sign of improvement, even if many families are choosing charter schools.

Levy and Merrow are also too quick to dismiss a substantial rise in DCPS graduation rates under Rhee and Henderson, though they are on firmer ground when they write that, “[e]ven if the graduation rate increase is real,” we should wonder about the value of DCPS diplomas when many students graduate from the school system unable to demonstrate a grasp of rigorous academic content.

Finally, and most importantly, student achievement is up over the past decade in the District of Columbia. Trying to establish causal relationships between student test scores and education policies is always a fraught exercise, and some of the reforms I write about have been in place only for a short time. But we know that the proportion of students scoring “proficient” or above on the rigorous and independent NAEP more than doubled, from 14 percent in 2007 to 30 percent in 2015, in fourth-grade reading and from 10 percent to 33 percent in fourth-grade math, bringing Washington up to the middle of the pack of urban school districts at that grade level, while the city’s black students largely closed gaps with African American students nationwide. As Merrow has written elsewhere, D.C.’s NAEP scores “have improved faster than any other urban district’s” since Rhee arrived in Washington. The next round of NAEP scores, which will start to reflect DCPS’s latest reforms, are likely to be released in early 2018.

But Levy and Merrow disparage the gains, claiming that they are the result of more affluent white families enrolling their children in Washington’s public schools. Most people would consider that a sign of growing confidence in DCPS. Further, the studies I consulted in the course of my research dispute Levy and Merrow’s claim. One, published by a consortium of major research universities, concluded that “overall, even when accounting for changes in student demographics, test scores in the District have improved substantially, especially in math.”

Another, by the District of Columbia’s Office of Revenue Analysis, reported that “NAEP results have increased not just across race, but across all subgroups such as gender, disability status and ELL [English language learner] status.” Levy and Merrow claim that a third study, the 2015 National Research Council analysis, “pointed out that most of the recent academic gain was likely the result of more affluent families . . . enrolling their children in public schools.” But the study doesn’t say that. Instead, it includes changing demographics among a number of factors behind DCPS’s improving academic results. Read it and see for yourself.

I wrote in my article that “student achievement has begun a long climb toward respectability” in the District of Columbia, but that DCPS “still has a long way to go.” A key challenge, I noted, is that achievement levels among Hispanic and black students, who together make up 82 percent of enrollment, lag badly behind their white peers. Levy and Merrow criticize me for giving these achievement gaps short shrift—only to use a statistic from my story about the performance divide to make their point that the problem is severe. The plight of Washington’s many impoverished students of color is indeed severe, especially in the half-dozen or so high schools serving largely impoverished student bodies in the eastern half of the city. New Chancellor Antwan Wilson should make these schools a high priority.

No school system can simply wave a wand and overcome the impact of poverty on the students in such schools. But the research I did for my article suggests that by overhauling its teaching corps, its instructional materials, and teachers’ daily lives in schools, DCPS has given its students a far better chance than they had before.

What would Levy and Merrow have done differently? “The schools of the twenty-first century must operate on different principles,” they tell us. “[S]tudents are the workers, and their work product is knowledge. This approach seeks to know about each child not ‘How smart are you?’ but, rather, ‘How are you smart?’ ” That’s it? Fortunately, they weren’t given the responsibility of rebuilding one of the nation’s most troubled school systems in 2007.

Our ideas can save democracy... But we need your help! Donate Now!

John Merrow and Mary Levy, with a reply by Tom Toch

John Merrow recently retired after forty-one years covering education for the PBS NewsHour and NPR. His new book is Addicted to Reform: A 12-Step Program to Rescue Public Education. Mary Levy is an independent education analyst and the former director of the Public Education Reform Project of the Washington Lawyers Committee for Civil Rights. Thomas Toch is director of FutureEd, an independent think tank at Georgetown University’s McCourt School of Public Policy.