Charter Schools and Their Enemies

PREFACE

In a sense, this story began back in the early 1970s, at a gathering of various conservative and neoconservative intellectuals, hosted by Irving Kristol, then editor of a high-quality quarterly publication called The Public Interest.

After a round of convivial recollections from those present about how we had begun our careers on the political left or, as in my case, the far left as a Marxist, Irving raised a very serious question about how some way could be found to improve the substandard educational levels of most black schoolchildren. At that point I said something like, “You are talking as if good education for black children is something that has never happened before, and that has to be created from scratch.”

This immediately caught his attention, and he asked me to tell him where this had happened, and how. I gave him a brief sketch of the history of all-black Dunbar High School in Washington, D.C., during the era from 1870 to 1955, and the various achievements of its graduates in elite colleges during that era, as well as in careers that led many of them to pioneer as the first black federal judge, the first black general, and the first black Cabinet member, among other distinctions.

His interest very much aroused, Irving urged me to research and write about this, and volunteered to finance the research. Out of this came an article titled, “Black Excellence: The Case of Dunbar High School,” which appeared in the Spring 1974 issue of The Public Interest. Two years later, I wrote another article for The Public Interest about a number of successful black schools, in various parts of the country, titled “Patterns of Black Excellence.”

If I thought that, amid all the research and writings about failing black schools, many scholars and policy-makers would be interested in black schools that succeeded, I was sadly mistaken. Many scholars and policy-makers already had their own explanations for the failures of black schools, and their own “solutions” for that problem. What I had written was, to them, at best a passing distraction, if not something that needed to be discredited, so that they could get on with promoting their own prescriptions, policies and programs.

Chief Justice Earl Warren had already declared racially separate schools to be “inherently unequal” in the Supreme Court’s landmark Brown v. Board of Education decision in 1954, so racial segregation was the prevailing explanation of substandard black educational achievements.

The fact that all-black Dunbar High School was only about a mile away from the Supreme Court where the Chief Justice made his historic pronouncement, and that Dunbar, at that time, sent a higher proportion of its graduates on to college than any white public high school in the city,¹ was a fact that was probably unknown to those crusading for racial “integration” in the schools, and that fact probably would not have made any difference to them, even if they had known it.

Many people had already made up their minds, and did not want to be confused by facts. Years of mandatory busing of black children to white schools, in order to achieve racial “integration” was the logical corollary of what Chief Justice Warren had said, though the Brown v. Board of Education decision did not itself prescribe mandatory busing. The busing crusade produced heated controversies, bitter racial polarization and dangerous confrontations in the streets, with schoolchildren caught in the middle of it all—but little, if any, net benefit to the educational levels of black children.

Eventually, the busing crusade faded in futility. But something very different later appeared on the horizon—the idea that low-income parents should be allowed to choose where their children went to school, just as high-income parents already could, by sending their children to private schools if the local public school was unsatisfactory. Extending choice to parents in general could be done in a variety of ways, including vouchers that could cover tuition at low-cost, private schools such as some Catholic parochial schools.

That was just one option among many. There were also magnet schools, homeschooling and tuition tax credits, for example. Eventually, one of the most strikingly successful kinds of schools that emerged from this experimentation was the charter school—a special public school freed from some of the rigidities of the regular public schools, and allowed to receive government financial support only so long as its students’ educational outcomes met various educational criteria.

Not all charter schools turned out to be successful, just as not all traditional public schools turned out to be successful—or all failures, for that matter. But particular charter schools, and especially some particular networks of charter schools, located in low-income black and Hispanic neighborhoods, achieved educational results not only far above the levels achieved by most public schools in those neighborhoods, but sometimes even higher educational results than those in most schools located in affluent white neighborhoods.

No one expected that. Anyone who might have predicted such an outcome beforehand would have been considered to be hopelessly unrealistic.

This story might seem to have had a happy ending—at least for that fraction of minority students attending successful charter schools. But, in fact, even the most successful charter schools have been bitterly attacked by teachers unions, by politicians, by the civil rights establishment and assorted others. How can success be so unwelcome? It is apparently not unwelcome to parents of low-income minority students. In New York City alone, there are more than 50,000 children on waiting lists to get into charter schools.² Yet New York’s mayor has announced an end to the expansion of charter schools and threatened restrictions on those already functioning. It is much the same story in California—and in many other places in between.

Understanding why and how educational success has been such unwelcome news to so many people and institutions is the purpose of this book. With growing political threats to charter schools across the country, the stakes could not be higher for poor and minority youngsters, for whom a good education is their biggest opportunity for a better life. That in itself is enough to make this a story well worth understanding by all people of good will, despite whatever other differences they might have.

Thomas Sowell

The Hoover Institution

Stanford University

Chapter 1

COMPARISONS AND COMPARABILITY

Depending on who you read or listen to, charter schools are either a striking success¹ or a “failed and damaging experiment”²—or even just “fads.”³ With all the voluminous educational statistics available, it might seem strange that such extremely different conclusions, and controversies arising from these differences, should exist and persist. Nevertheless, these controversies have continued to rage for years, with growing intensity, as charter schools have expanded from a barely noticeable part of the educational scene when they began in the 1990s to thousands of schools with millions of students today.

Public charter schools are public schools not created by the existing government education authorities, but by some private groups who gain government approval by meeting various preconditions set by authorizing agencies.⁴ These agencies issue charters enabling these schools to operate as public schools eligible for taxpayer money and to enroll public school students who apply.

By allowing more autonomy and flexibility in public charter schools than in the more tightly controlled traditional public schools, it was hoped that new educational policies and practices that emerge from this experiment might produce some better educational results. In that case, traditional public schools would have these new policies and practices available to use if they chose to, thereby benefitting the much larger number of students in the traditional public school sector. If, however, a charter school has educational outcomes that fail to satisfy the authorities, those authorities can revoke its charter and end its access to taxpayer money and public school students.

One important difference, however, is that students are not assigned to go to public charter schools, as they are assigned to attend particular traditional public schools. Those students whose parents want them to go to particular charter schools can seek admission to those charter schools, usually by entering a lottery. Choosing students by lottery—rather than by their ability or their educational track record—is supposed to keep the students in the two kinds of schools more or less comparable, so as to keep the experiment valid and its conclusions applicable to public schools in general.

One major complication in studies comparing public charter schools with traditional public schools is that the racial, ethnic, and socioeconomic backgrounds of students in the charter schools as a whole turn out to be very different from those of students in traditional public schools as a whole.

COMPARABLE STUDENTS

Nationwide, white students plus Asian students are a majority of the students in traditional public schools, while black students plus Hispanic students are a majority of the students in charter schools, which are often located in low-income minority communities.⁵ On a wide range of educational tests, over the years white and Asian students as a whole have scored significantly higher than black and Hispanic students as a whole.⁶ Therefore comparisons of charter school and traditional public school outcomes on various tests are a problem, because their respective students are from groups with a long history of different educational results. There is also a long history of different educational results with children from low-income families and high-income families.

Under these circumstances, it can be hard to know how much of whatever differences there may be in educational outcomes, as between charter schools and traditional public schools, are due to the schools themselves and how much are due to their different mix of students from different ethnic and socioeconomic backgrounds.

A striking example of how racial or ethnic differences among students can make it hard to determine the effectiveness of different schools—whether in terms of charter schools or in other contexts—is a study of educational test score differences among the 50 states. Students in Iowa scored higher on those tests than students in Texas. But whites in Texas scored higher than whites in Iowa; blacks in Texas scored higher than blacks in Iowa; Asians in Texas scored higher than Asians in Iowa; and Hispanics in Texas scored higher than Hispanics in Iowa.⁷ How then could Iowa students as a whole have scored higher than Texas students as a whole? Simply because “Iowa’s student population is predominantly white”⁸ and students in Texas include far more minority students, mostly low-income minorities.

While gross statistics might suggest that Iowa had better schools than Texas, an ethnic breakdown of the population taking those tests would suggest the direct opposite. For similar reasons, comparing educational outcomes in charter schools as a whole with educational outcomes in traditional public schools as a whole can be like comparing apples and oranges—unless there is some way to compare particular schools’ educational results when educating truly comparable students.

Since such comparability is simply not there in gross statistical comparisons of public charter schools as a whole with other public schools as a whole, the approach here will be to compare individual charter schools with individual traditional public schools that are as similar as possible. Among the wide variety of statistics available on educational test results in charter schools and traditional public schools, the ones given the greatest weight here will be statistics comparing students in particular schools meeting all three of the following criteria:

Schools meeting all three requirements simultaneously are by no means common. But, if our goal is to compare educational results among truly comparable students in truly comparable circumstances, whether those students are in charter schools or in traditional public schools, then this may be as close as we can come to achieving that.

Uncommon as it may be to find large numbers of such situations in a given community, New York City is exceptional in having a substantial number of charter schools and traditional public schools meeting all three requirements. In school year 2017–2018, there were more than 23,000 New York City students in particular classes meeting all these requirements in these particular schools.¹⁰ So New York City has a substantial sample of ethnically and socioeconomically comparable students whose educational outcomes can be compared.

THE DATA

Each school year, the New York State Education Department gives the same tests in “English Language Arts” and in mathematics to public school students—whether in public charter schools or in traditional public schools—in grades 3 through 8. So it is possible to make comparisons of students’ results on these tests in the same grade levels in particular charter schools with particular traditional public schools located in the same buildings. The New York State Education Department publishes not only aggregate test scores of classes in these schools but also the ethnic breakdown of the students and the percentage of them who meet its definition of “economically disadvantaged.”

That still leaves the question of how to select which of the innumerable pairings of classes to examine. If the pairings are chosen by simply cherry-picking examples, all the efforts to achieve comparability will have been wasted, since different people can obviously choose different examples.

One viable option is to simply make available all the data from all of the classes in New York City where a charter school has been housed in the same building with one or more traditional public schools which have some classes at the same grade levels, and students with a similar racial or ethnic background. That is done here in the Appendix. But while it is a viable option to make available all test score data, demographic data, and socioeconomic data for students in these situations, it is not a viable option to discuss all these data individually, without expanding the study to the dimensions of an encyclopedia. Moreover, charter schools differ among themselves, just as traditional public schools do, and these differences also require discussion and analysis.

Selecting which charter schools to examine in detail by some principle, as distinguished from arbitrary cherry-picking, can be done in a number of ways. The way chosen here is to examine those particular charter school networks with multiple schools having classes in the largest number of buildings in New York City where they are housed with one or more traditional public schools whose grade levels coincide. Here the sample chosen for detailed study in Chapter 2 are all charter school networks with students in five or more buildings in New York City that they share with traditional public schools having students at the same grade levels.

For the sake of ethnic comparability, the paired schools in our sample must also meet the criterion of having a majority of their students who are either black and/or Hispanic. A finer ethnic breakdown for each school can be found in Appendix II.

Such data provide separate samples from different charter school networks, and from different school locations within each network. As a result, the statistical influence of the peculiarities of any particular school or any particular neighborhood on the data can be reduced or at least recognized.

People who would prefer some other method of choosing samples to examine in detail are free to make their own selection from the voluminous data in the Appendix. All these data are from the New York State Education Department, and the definitions used in the tables are their definitions.

By choosing to examine in some detail those charter school networks with classes located in five or more buildings in New York City, a large amount of data can be examined from a small number of charter school networks. In this case, there turned out to be five charter school networks in New York City that met the three specified requirements in school year 2017–2018, and had classes housed in five or more buildings with traditional public schools having classes at the same grade levels. These networks were the KIPP (Knowledge Is Power Program) charter schools, and charter schools in the Success Academy, Explore Schools, Uncommon Schools and Achievement First networks.

After examining the performances of these particular charter school networks in some detail in Chapter 2, there will be a more summary examination there of the performances of all charter schools in New York City that met the same three criteria for inclusion in the citywide sample.

Chapter 2

CHARTER SCHOOL RESULTS

While our main concern is finding out what educational outcome differences there are between students in public charter schools and students in traditional public schools, the detailed data in our sample also reveal differences in test scores between different charter school networks, between different schools in the same networks, between classes in the same schools, as well as similar differences among traditional public schools. All these differences can be found in the tables in this chapter, by those who are interested, even if not all these things are discussed in the text.

Data on the ethnic makeup of charter school students and traditional public school students paired with them in the same buildings are available in Appendix II, and are summarized in passing in this chapter. In both kinds of schools, these are ethnic data for students in the specific classes being compared, not data on the ethnic makeup of students in the entire buildings in which they are housed. Our samples are defined by the classes whose test scores are being compared, not by all the students in the buildings.

Two tests given annually by the New York State Education Department, to both public charter school students and students in traditional public schools, are officially designated as the English Language Arts test and the Mathematics test.

The students’ scores on these statewide tests are broken down into four categories by the New York State Education Department. The lowest test scores are officially defined as being in Level 1, and the highest test scores are defined as being in Level 4. Students who score in Level 3 are designated as being “proficient,” according to the standards for whatever grade they are in, and those whose scores are in Level 4 are designated as being above “proficient” for that grade. These definitions are repeated under each table of statistics showing test score results. The main point here is simply that Level 1 is at the bottom and Level 4 is at the top.

“Proficient”—Level 3—is a crucial measure. While students who fall below that level are likely to be promoted to the next grade anyway, in many or most traditional public schools, their prospects of mastering those subjects in higher grades that build on what was taught in the same subjects in lower grades—mathematics being a clear example—are obviously not good. That is especially so if they score in the bottom category, Level 1, two levels below “proficient.” Therefore statistics on test scores in Level 1 are also crucial, and will also be a special focus here. To score two levels below “proficient” in arithmetic makes it unlikely to be able to master algebra in later years. Cumulative deficits can be extremely hard to overcome, even by conscientious and intelligent youngsters.

In a world where higher mathematics is required in many professions—not just for scientists, engineers or statisticians, but increasingly also for economists, psychologists, sociologists and others*—an inability to master mathematics means that doors of opportunity into a wide range of professions are silently closing in the background as children go through elementary school without achieving proficiency in arithmetic. Having children talking in school about how they are going to become doctors or pilots, when they have not mastered fractions or decimals, is a cruel hoax—as they can discover later in life as adults, when it is too late.

KIPP CHARTER SCHOOLS

The KIPP charter schools are the largest non-profit network of charter schools in the country. The first of KIPP’s more than 200 schools, now scattered from coast to coast, began in Houston in 1994 and the second, a year later, in New York City’s South Bronx. Both schools serve predominantly minority youngsters from low-income families, as do other schools in the KIPP network.

In New York City, there were 11 KIPP charter schools in school year 2017–2018, including 5 located in the same buildings with one or more traditional public schools serving the same community, and having some classes at the same grade level in both kinds of schools. In each of these 5 KIPP charter schools, at least 95 percent of the students in our sample were either black or Hispanic in 2017–2018. This was also true of the ethnic breakdown in the traditional public schools housed in the same buildings.¹ Most of the students in both the KIPP charter schools and in the traditional public schools housed with them were classified as “economically disadvantaged” by the New York State Education Department.²

High Scores in English

In school year 2017–2018 a majority of KIPP charter school students scored at Level 3 (“proficient”) or above on the English Language Arts test in 10 of their 14 grade levels in the five buildings they shared with students in traditional public schools.³ A majority of the traditional public school children in these same five buildings scored at Level 3 (“proficient”) or above in just one of their 20 grade levels. Some of these buildings contained more than one traditional public school, which is why there were more grade levels for traditional public school students than for KIPP charter school students. Details are shown in Table 1A.