Charter Conundrum

C harter schools probably will not settle the education wars, but they may provide an armistice. Conservative privatizers see charter schools as a next-best alternative to voucher plans, which have now lost political momentum; progressive educators, on the other hand, see charters as places where they can implement long-sought reforms, free from constraints imposed by rule-bound school bureaucracies. Each side hopes to exploit charter schools' disarmingly simple trade-that almost any group can get public funds to run almost any kind of school, provided they are "explicitly accountable" to the public for "improving student performance," in the U.S. Department of Education's words.

But there's the rub. The premise on which charter schools are based-that we can hold schools accountable for results-is a myth. As contemporary debates about national standards and testing show, there is no consensus about how to assess educational outcomes objectively. Nor does political accountability seem promising; charter schools never lose their charters for academic malpractice.

Yet almost all supporters of charter schools agree that accountability will be the fulcrum of their success. The "charter" is a contract between those who operate a school and the school district (or other government entity) that issues the charter: if you prove you can boost student achievement, you can get there any way you like. Former California State Senator Gary Hart (no relation to the former Colorado senator), an early advocate of charter schools and sponsor of the California policy, put it this way: "the tradeoff has always been outcomes versus deregulation. And if we can't demonstrate the outcomes, we're not entitled to the deregulation."

But if charter schools can't be held accountable for results, it's not clear how the movement will transform public education. In the absence of measurable outcome standards, many of the claims (both praise and criticism) about charter schools are hyperbole. Will charters eventually reveal common ground between left and right, on which future improvements in America's public school system might be based? Or are charters merely recasting the terms on which the big battles will once again be fought? These questions are far from settled, but the early development of charter schools at least offers us some basis for judgment.

WHAT ARE CHARTER SCHOOLS?

The first charter schools were authorized in Minnesota in 1991 by the same group of school critics who pioneered the first statewide school choice programs a decade earlier [see Ross Corson's "Choice Ironies: Open Enrollment in Minnesota," TAP, Fall 1990]. In 1998, about 700 charter schools operate in 20 states, enrolling about 170,000 students-less than half of 1 percent of total public school enrollment.

President Clinton has called for the establishment of 3,000 charter schools by 2000 (still a tiny fraction of the nation's 85,000 public schools), and the federal government now provides planning grants to charter operators. "[L]et teachers form new schools with a charter they can keep only if they do a good job," Clinton said in his 1995 State of the Union address. Other strong supporters of charter schools include the Democratic Leadership Council and the Hudson Institute, a conservative think tank whose education programs are spearheaded by voucher advocate Chester Finn, previously the Reagan administration's assistant secretary of education.

Proponents herald the charter school movement as the salvation of public education, and the schools' public character is reflected in their admission policies. Charters generally must be open to all students who apply. (The federal government and many states require oversubscribed charters to choose students by lottery, although many ignore this policy and substitute a first-come-first-served system.) This formal requirement does not prevent charters from influencing student selection more carefully than neighborhood public schools do. Charters can assure a compatible student body by their targeted recruiting practices, counseling, and advertised philosophies. Of course, neighborhood public schools also reflect the socioeconomic characteristics of the stratified and segregated neighborhoods in which they are located, so charter schools may be only marginally more homogeneous.

×

There are many kinds of charter schools. Christopher Whittle's Edison Project, having failed at trying to operate for-profit independent schools, now runs charters on contract with public agencies. There are Montessori charters and "back-to-basics" charters. There are charters with E. D. Hirsch's Core Knowledge or Mortimer Adler's Paideia curricula, and there are "online" charters whose students communicate with tutors by computer and have no formal classes to attend or school building to sit in. Some states exempt charter teachers from collective bargaining protection; other charter schools are sponsored by teacher unions. There are charters teaching creationism whose affiliation with conservative fundamentalist Christian groups is barely disguised; charters with Afrocentric curricula; and charters sponsored by civil rights activists and organizations like the National Council of La Raza. There are inner-city bilingual charter schools and middle-class home-schoolers who have found charters to be a way to tap into public funds.

SHAM ACCOUNTABILITY

How do we evaluate these diverse charter schools-both against conventional public schools and against each other? The truth is that despite years of hyperventilated rhetoric about declining student performance, falling test scores, and failing regular schools, we have little reliable achievement data with which to compare schools. American students are certainly tested often with standardized measures. Some states administer official standardized tests to all students; in others, commercial test publishers sell exams that districts or schools may or may not use. The scores are widely publicized, but they're overemphasized and misleading, for these reasons:

• Student achievement is only partially affected by school effectiveness. Most of the variation in student performance is determined by the social and economic background of students. Adjusting reported outcomes for differences in such characteristics is notoriously difficult; there is always the possibility that a school with low scores and disadvantaged students might be a better school than one with higher scores and more advantaged students.
• Different tests may emphasize different skills. Student scores on a math test that rewards calculation prowess cannot easily be compared to scores of students on a test that rewards problem solving. This makes it impossible to "equate" scores from one test to another, unless we want national standardized testing-which raises its own problems [see Peter Schrag's "New Page, Old Lesson: Why Educational Standards Fail the Political Test," TAP, March-April 1998].
• Testing is expensive, and when test publishers and schools cut corners to reduce these expenses, they render tests less valid. For example, commercial tests are updated only about once every five years. As a consequence, scores of students at any grade level often increase yearly, as teachers learn to teach to the test.
• Test conditions are not standardized from place to place, making comparisons impossible. Test scores at Los Angeles's highly regarded Vaughn charter school (recipient of California's "Distinguished School" award) fell after the school gained autonomy from its district. The charter school principal, committed to full inclusion for all children, had tested both children with learning disabilities and children who had recently transitioned to English-language classes from the bilingual program. But before the school became a charter, neither of these groups was included in test score reports. Is the charter school now doing a better or worse job than the regular school it replaced? Nobody can say for sure. Nationally, there are no generally accepted protocols for making decisions about which students should be included in and excluded from testing.
• Our lack of sophistication in test score interpretation leads to illogical standards not only for charters but for all schools. A typical charter, for example, makes a promise that "all children will perform at grade level" in reading or math. But "grade level" performance is, by definition, a national average of students who perform both below and above a typical level for children their age. No matter how high we make our standard, there will always be a distribution of scores around it. To promise that all children will meet a standard must mean that the standard will be set so low that few children will be challenged by it.

B ecause of the difficulty of meaningfully holding schools accountable, not a single charter has been revoked for measured academic underperformance anywhere in the country since charter schools began in 1991. One school, a K-12 American Indian charter in Redwood Falls, Minnesota, was denied permission to continue offering an academic program in the high school grades (but permitted to continue its elementary program) after the Lower Sioux Indian Community Council complained to the Minnesota Board of Education that the upper grades were functioning more like a "teen drop-in center" than a regular school. Several other charters elsewhere in the country have been revoked or restricted for financial mismanagement or for violation of church-state rules. But none have been revoked for academic failure. While many charter schools may be academically successful, it is improbable that all are.

The highly publicized success of the Edison Project's Renaissance Charter School in Boston [see Peggy Farber, "

Boston: Renaissance Charter School

"] is one illustration of the inadequacy of our accountability measures. In October 1995, at the beginning of its charter, Renaissance administered the Metropolitan Achievement Test (a nationally normed exam published and sold by Harcourt Brace). Third graders' median grade equivalent (G.E.) reading score was then 2.7-that is, in October they were reading at the average level, nationally, of second graders in the seventh month of the school year (March), meaning that Renaissance third graders were then about five school months behind the national average.

In October 1997, Renaissance again administered the MAT, and fifth graders (comparable students, two years later) had median grade equivalents of 5.6-they were now four months ahead of their peers in the second month of fifth grade. From five months behind to four months ahead is an impressive two-year gain, and the Edison Project and other charter proponents use these data to advertise charter schools' success.

But the data are less conclusive than they appear. Not only does the MAT suffer from the problems of standardized tests described earlier, but the test score data have no controls for student turnover. Renaissance, like all charters, is a "choice" school. Dissatisfied students will leave, while students more motivated to succeed will be accepted from the waiting list. If, from 1995 to 1997, less successful students left and more accomplished students entered, the apparent gains in reading scores would be misleading. If, on the other hand, students on the waiting list tended to be those who were more likely to be failing at their prior schools, then the apparent gains could understate Renaissance's success. From the school's reported average scores, we can't know whether or how much they are biased by either of these conditions.

Even if data show that Renaissance is teaching reading effectively, they can't support conclusions about whether the school's charter status contributed to this success. Renaissance teaches reading by contracting for a program called "Success for All," designed and sold by a team led by Robert Slavin of Johns Hopkins University. "Success for All" consists of a balance between phonics and whole language approaches to reading, one-on-one tutoring by certified teachers for children having reading difficulty, extensive teacher training, and family support services based in the school. About 750 low-income public schools in 36 states nationwide have purchased the program, mostly using Title I federal compensatory education funds to do so. "Success for All" is probably the most effective reading program available today for disadvantaged youngsters; a national study comparing "Success for All" schools with matched comparison schools that did not use the program found that fifth-grade reading in "Success for All" schools was a grade equivalent higher than in the comparison schools. We have no data on how Renaissance's reading progress compares to progress in the hundreds of regular low-income public schools that also use this program. Without such data, we can't say whether the governance characteristics of Renaissance (independence from the Boston school district, for example, and profit incentive for Edison) played any part in its success, or if it did, how much of a part.

T he difficulty of ensuring accountability is further illustrated by Classical Academy, a Colorado Springs charter established by members of the conservative Christian group Focus on the Family. According to William Celis, who visited the school in reporting for this article, the school's founders have created a curriculum that attempts to combine the "Core Knowledge" program designed by University of Virginia professor and author E. D. Hirsch with a "character building" emphasis on "loyalty, tolerance, courage, and perseverance." The school's mostly middle-class student body includes not only some previous public school attendees, but former home-schoolers as well. The school's principal was previously Focus on the Family's national educational policy director.

This Classical Academy charter, issued by the state of Colorado, requires the academy's test scores to meet or exceed the average for all students in Colorado Springs. To establish baseline data, the school administered the Iowa Test of Basic Skills (ITBS, a commercial test published by Houghton Mifflin) in September of 1997, its first year of operation. It found that its students initially had scores below the district average; thus, Classical Academy will have to make up ground in order to meet the terms of its commitment.

Though typical of the state's charters, this requirement is very curious. A Colorado charter whose students entered with scores higher than those typical for its district could see its scores drop from ineffective teaching, yet still meet the state's requirements; another whose students began far below the district average could be unusually effective, but not effective enough to meet the terms of the charter law. Nor do these inconsistencies reflect the fact that many districts have a wide variation in student family characteristics. Classical's local school district is mostly white and affluent, with some 13,000 students in 19 different schools. But even in Colorado Springs, minority students (about 10 percent of the population) and students who receive subsidized lunches (about 5 percent) are not evenly distributed. Even if we accept ITBS scores as a reasonable standard by which the Classical Academy should be judged, to assess the school's performance intelligently we need to know how its scores compare to those at demographically comparable schools, not to those of the district as a whole.

Consider also the Bowling Green Elementary School, a regular public school until 1993 when its teachers petitioned the Sacramento school district for autonomy in return for guaranteed results [see Susanna Cooper, "

Sacramento: Bowling Green Elementary

"]. The teachers' application noted that "the existing educational program is not working for our students" because Bowling Green's reading scores, on average, were at the twentieth percentile of the national norms established by the Comprehensive Test of Basic Skills (published by McGraw-Hill), compared to a Sacramento district-wide average of the thirty-second percentile. "We want to show that public education can work when managed differently," the Bowling Green charter application states. So it commits the school to report, each year, the number of students who reached a proficiency standard on California's statewide assessment, and commits each of the school's sixth graders to accomplish all five of these goals: know algebra; be fluent in either English or Spanish and conversational in the other; walk, run, or use a wheelchair to cover a mile in ten minutes; read and interpret a variety of materials; and research and then provide a community service. The charter pledged to use both state tests and a series of portfolio demonstrations by students to assess progress in these domains, and, in a burst of enthusiasm, promised to train parents to meet each of these goals as well, using newly hired community workers not only to recruit parents to assist the school, but also to instruct them.

After granting the charter, however, California canceled its statewide assessment in a controversy about whether the test gave too little emphasis to basic skills and violated students' privacy by asking them to write essays reflecting personal feelings and experiences. In the last six years, the Sacramento district has switched testing instruments three times, and the different tests don't report scores for the same grade levels. In 1997 the district adopted a locally developed and administered test; this year it began to use Houghton Mifflin's ITBS. Nobody can now say with any assurance how scores on the new test relate to scores on the previously used California state test or the district's locally developed test. As Susanna Cooper reports, Bowling Green shows its students rising in percentile rank relative to other district schools; however, this may be because recent tests are better aligned with Bowling Green's curriculum, while past tests were better aligned with the curricular focus of other schools.

Meanwhile, Bowling Green's teachers have abandoned the charter's other lofty goals because, according to the principal, "We came to the realization that we couldn't do it all. At that point reading became the main focus. As we refined it we became more like a traditional school." In 1996, the school board voted to renew the charter, satisfied with the school's apparently improved reading scores relative to other district schools-notwithstanding the charter's earlier promise that it would be revoked for "failure to meet or pursue any of the pupil outcomes identified" [emphasis added].

Finally, the Minnesota New Country School [see Ross Corson, "

Le Seuer-Henderson: Minnesota New Country School

"] presents a different dilemma. Like almost all charters, the MNCS contract provides that the school "must design its programs to at least meet the outcomes adopted by the State Board of Education" and that the state will cancel the charter for "failure to meet the requirements for pupil performance." But like most state exams, Minnesota's minimum competency exam is so minimal that New Country students, coming from relatively advantaged and educationally motivated families, could probably pass it if they never went to school at all.

The school's founders and teachers wanted a charter school precisely because they rejected standardized tests as the principal means of evaluating student work. Thus the MNCS charter commits it to "create different and innovative forms of measuring outcomes," with each student developing an "individual learning plan" negotiated with the school's teachers. Academic achievement at New Country is assessed primarily from project portfolios, and once every two months students make public presentations of their projects at the local shopping mall to an audience of parents and interested citizens.

Most thoughtful educators applaud these approaches, and many regular public schools now attempt portfolio assessment as well. (Their attempts are constrained by critics' demands that teachers orient instruction to the narrow purpose of boosting scores on standardized tests that emphasize multiple-choice questions.) New Country School leaders can perhaps be proud of the school's innovations, but by de-emphasizing standardized testing, their approach can offer no hard data by which we can judge whether their status as an "outcome-based school" (the Minnesota term for "charters") does, in fact, boost outcomes.

I n the absence of hard data, charter school proponents rely on anecdotes to show that charter schools are superior. In the worst cases, anecdotes are disguised as data. For example, in a recent widely publicized report, the Hudson Institute surveyed charter school participants to determine levels of satisfaction, and announced results like the fact that "62.5% of black students say they like their teachers better [than in their former regular schools] and 56.1% say they are more interested in their school work." Of all students, "60.7% say that their charter teachers are 'better' than teachers at their previous school." Statistics like these lead Finn and his colleagues to conclude "that there is particular satisfaction among students who have left traditional public schools and that student satisfaction crosses racial and ethnic lines"; that "[c]harter schools are havens for children who had bad educational experiences elsewhere"; and that "[t]hese satisfaction levels for charter schools stand out in a time when there is growing dissatisfaction with traditional public schools. . . . There seems to be a consensus among all primary constituents that charter schools are living up to their expectations and delivering a high-quality product (or at least improving upon the alternative)."

It's too bad the Hudson Institute's analysts are apparently less numerate than they expect charter school graduates to become. If there were no difference in quality between charter schools and regular schools, and if student satisfaction were randomly distributed, we would expect about 50 percent of charter school students to be more satisfied with their teachers than they were with teachers in their prior public schools, and about 50 percent of them to be less satisfied. Because charter school parents and students have chosen to be there, we should also expect surveys to show a predisposition in favor of the charter school, even with no difference in quality. Thus, it is somewhat underwhelming that only 61 percent of students say their charter school teachers are better than their previous public school teachers. The statistic certainly does not provide a sound basis for the ringing endorsement of charter schools that Finn and his colleagues find in it. In short, in the absence of hard data, asking charter school participants whether they are satisfied does not meet the accountability standard charter school proponents promised-nor is it likely to give us a basis for evaluating charters' validity as an educational reform.

Charters leave us with the dispersal of public funds to educational experimenters, with no meaningful accountability. This does not mean charter schools are ripping off the taxpayers, nor that charters are worse than regular schools, which are also largely unaccountable for results. Some charters may be just as or more effective than most regular public schools, and some may be less so. But what we hear about charter schools today is little more than anecdotal puffery and phony statistics, the same sort of anecdotes and flawed data so often used to condemn-or, less frequently, to praise-regular public schools.

"CREAMING"

Accountability is further confounded by the possibility of "creaming": even if schools of choice are technically open to all students, those who choose to attend will be more motivated, more accomplished, and more sophisticated than those who remain in public schools. If so, choices will concentrate the advantaged in charter schools and leave schools from which they came with lower average academic and socioeconomic characteristics. These concentrations will result in inferior educations for students remaining in regular schools. Studies of existing choice programs confirm that this segregation does in fact occur. For this reason, many public school choice programs, like urban magnet school programs, or like admission plans for all schools in places like Cambridge and Boston, have limited choice tightly, imposing rigid racial and ethnic quotas on student admissions.

Charter schools are generally not permitted to give admission tests (an exception might be a charter school focused on a theme like music), but are also not required to impose quotas to assure a representative student body. Are they, then, likely to become a "creaming" or segregating force in American public education? Because charters are so new and encompass such a wide range of school types, this is still an impossible question to answer with certainty, but we have hints.

Charter supporters emphasize that existing charters seem to enroll a larger proportion of racial minority and poor students than do comparable public schools in the states where these charters are located. At the end of 1996, about 14 percent of all charter school enrollment was African-American and 25 percent was Hispanic; 20 percent of all charter schools had less than 20 percent white enrollment; 33 percent of charter school students were eligible for the federally subsidized lunch program. "One might suppose," Finn and his colleagues conclude, "that the 'creaming' allegation could now be laid to rest."

Not so fast. These data are partly driven by regulations of the sort that conservative charter proponents detest. California's charter law requires that priority be given to schools serving low-achieving students. Colorado mandates that one-quarter of the state's initial charters must be for schools primarily serving "at risk" students; Rhode Island requires that half its charters go to such schools. South Carolina mandates that the racial distribution of a charter school's students must parallel that of its district. (This requirement frustrated the efforts of a group of wealthy Hilton Head parents to charter a school, because they would have needed 40 percent black enrollment to be typical of Beaufort County.)

High percentages of minority students in charter schools are also partly the result of a considerable number of schools like Bowling Green, where teachers in an existing inner-city school have applied successfully to convert to charter status, gaining independence from regulation in return for a commitment to boost test scores. Minority percentages also reflect schools like Renaissance in Boston, where Edison determined that the credibility of its future efforts to create a proprietary education market might depend, at least initially, on its ability to boost test scores in an inner-city loss leader. If the data on charter schools excluded conversion charters like Bowling Green, or proprietary loss leaders like Renaissance, the race and poverty distribution of charter enrollment would be less impressive.

Forces tending toward creaming include laws like California's that do not require charters to offer school lunch programs, because not offering lunches effectively restricts the types of students who can apply. Schools like MNCS or Classical Academy were founded not to serve children who were failing in existing inner-city schools, but to pioneer alternative pedagogies for students whose dissatisfaction with regular schools was not based on lack of success. Even if charter applicants are no more advantaged or able than students in the schools from which they come, if applicants and their parents previously were most dissatisfied, and thus more likely to pressure regular schools for improvement, loss of these pressures could damage regular schools. If charter schools become an exit strategy for public school parents, regular schools could suffer from losing activist parents.

Special education presents another "creaming" danger. Charter school legislation prohibits schools from refusing to admit or serve a student with disabilities, no matter how severe. And all the special education rules-requirements to provide whatever services are needed, to include disabled children in regular programs wherever possible, and to design an individualized program for each disabled student-apply to charter schools as well. But despite these obligations, as Peggy Farber's report on Edison's Renaissance School shows, a charter school could effectively limit special education obligations by recruitment and counseling policies that might formally meet requirements but effectively discourage special education enrollment.

This year, after the events described in Farber's report, Edison hired a new Renaissance principal and recruited a respected special education specialist to make the school's "full inclusion" program work. Teachers are now better trained to handle difficult problems in the classroom, and they are assisted by a specially trained teacher on each floor. (And Kylee Jones, the student at the center of Farber's report, is now back at Renaissance School.) It is a credit to the Edison Project that it moved so effectively to address Renaissance's inadequacies. But charter advocates who believe deregulation is a panacea should note that it may have required the intervention of public regulators to get Edison to reform.

CHARTERS AND REGULAR SCHOOLS

Some schools receiving public funds under state charter provisions are indistinguishable from private schools. As of the end of 1996, about 10 percent of all charter schools were previously private schools that had received public charters and funds. In California and Arizona, groups of home-schoolers banded together and won funds to continue their private efforts at public expense. In Arizona, a Montessori school also converted to charter status, confident that its well-defined pedagogy would successfully filter out unwanted applicants and preserve its distinctive character.

At the other extreme, the most creative and innovative approaches of the best charter schools can also be found in the ongoing experimentation that takes place in public school systems without formal "charter" status. Some, like New York's "New Visions" high schools, are semi-autonomous, thematic public schools operated in partnerships with museums, parent groups, and a health care union, among other organizations. Boston's "pilot schools" are also not technically charters, but district policies and teacher contract provisions are waived to permit greater experimentation.

Charter advocates reasonably argue that were it not for the threat of chartering, New York and Boston might never have offered these options. On the other hand, new pedagogies like "Success for All" have spread through public school systems without any spur from the charter or other privatization movements, and ongoing public school innovation is already reflected in the classrooms of the tens of thousands of progressive public school teachers. There are few models of education carried out in charter schools that are not also being carried out in many public schools without formal autonomy. The accompanying report on Sacramento's Bowling Green charter, for example, notes the creative ways in which the school has used its Title I funds. But the 1994 reauthorization of the federal Elementary and Secondary Education Act now permits regular schools the same flexibility in utilizing compensatory money, so each dollar no longer must be tracked to an individual eligible poor child.

Of course, charter school teachers and the parents who support them may consider it advantageous to be in a community where their progressive education philosophies are incorporated in a school definition, not subject to the (sometimes grudging) tolerance of a principal or school board. And so some teachers and parents will continue to form charters that will increasingly look like regular public schools, even as public schools increasingly support experimentation with new pedagogies and practices.

Formal charter status will not likely become a dominant trend in public education. Most of the nation's nearly 3 million teachers and 100 million public school students and parents are not experimenters and innovators. This is true of any institution, and we would be foolish to expect schools to be otherwise. Poll after poll finds that while Americans have an ideological commitment to education reform and believe our schools are presently inadequate, most people also believe that the particular public schools their own children attend are doing just fine. The incentives for rapid growth of charter schools are not that strong.

LIMITS TO GROWTH

Charter schools that do form will increasingly limit their own independence as charter operators discover that the school bureaucracies they love to hate exist, at least in part, for good reason. This has been a recurring trend in a series of decentralization reforms enacted in American education over the last decade. "Shared decisionmaking" and "school-based management" reforms in public school systems have often foundered because teachers concluded they don't want to waste time and energy in committee meetings deciding which copy machine to purchase or what the school lunch menu should contain. Teaching was already a full-time job, and teachers inclined by inexperience to believe that administrators did nothing but collect paychecks have come to the realization that many bureaucrats play indispensable roles. The most successful school-based management programs have involved teachers in serious policymaking, but have retained the support staff to continue its administrative function.

But many charter school operators fail to make this distinction. In an initial burst of enthusiasm, teachers and parent volunteers in many charter schools will be willing to work 80 hours a week, clean bathrooms, and give up weekends and vacations to realize a vision of what schools can be. But this can get tiresome, as the accompanying description of Minnesota's New Country School suggests. Finn and his colleagues at the Hudson Institute have often blamed alleged failures of our public schools on "bureaucrats" and government control. But one of the most revealing passages in Hudson's recent encomium to charter schools acknowledges that the teachers, principals, and parents who start charter schools "are often dreamers and visionaries, less often veterans of meeting payrolls, balancing budgets. . . . We never expected to say this-critics as we generally are of school-system bureaucracies-but some charter schools are simply under-administered."

Yet another limitation to charter school growth is financial. Many of the more conservative advocates of choice (and now charters) have always claimed that public schools are inefficient and could deliver better instruction for less money. Notwithstanding this oft-repeated claim, however, charters find that better teaching may require more money, not less.

Most state charter laws require that charter schools be given the same per-pupil funding regular schools receive. But since charters don't generally get capital funds (although some states have begun to provide them) or access to public bonding authority, most of their start-up funds must be raised privately. Thus charter supporters can rightly claim that they do not really receive equal funding. But there are offsetting factors. One is that when elementary school charters receive a district's average per pupil funding, they actually get more funds than other elementary schools, because secondary schools are more expensive to operate and pull up the per pupil average. Another is that charter school teachers often must be more energetic than ordinary teachers; as a result, charter teachers are more likely to be young. This means that they fall lower on traditional teacher salary scales than the typical teacher does, and so a charter school that receives a district's average per pupil funding (boosted to a considerable extent by higher salaries paid to more senior teachers) will be able to hire more teachers than traditional schools can. This is not insignificant. The typical school district compensation schedule has a two-to-one ratio (or greater) of top to bottom teacher salaries. A charter school that hires a predominantly young teaching staff can reduce class sizes considerably without paying teachers any less than they would receive in regular schools. This is not a replicable strategy for American education generally.

It is impossible to say whether, on balance, charter schools now receive funding comparable to regular school levels. But the fact is, charter schools' enriched programs are often more expensive than public school programs and so require greater funding. In some cases, charter schools can implement their visions only by cutting desirable, but less essential, programs. The decision of Bowling Green to fund smaller class sizes in part by eliminating a physical education teacher (while regular teachers give up their rest periods to supervise relatively unstructured playground exercise) is one example. In other cases, charter schools may substitute volunteer labor to do tasks for which regular schools must pay: a bilingual charter school in Oakland, California, keeps class sizes lower by doing without a janitor and relying on parents for custodial work. And in most cases, charter schools supplement their allotment of public funds by strenuous private fundraising.

N one of these solutions can endure. Once the flush of excitement about charters' unproven (and unprovable) higher test scores abates, Americans are unlikely to abandon their long-held conviction that well-rounded schooling includes physical education, art, and music as well as math and reading. Volunteerism can be relied upon in the initial stages of an experiment, but if a reform becomes institutionalized, volunteers must give way to employees. And corporations and foundations may willingly (or, if you ask charter school fundraisers, reluctantly) subsidize experimental schools, but if they are going to end up subsidizing all schools, they may as well do so through the tax system. (If President Clinton's proposal to fund national primary school class size reduction to 18 is enacted, many of the programs charter school proponents advocate could more easily be implemented in regular schools.) Our unwillingness, thus far, to fund education at levels necessary to produce the achievement we desire does not bode well for the replicability of charter school experimentation.

But there is also no evidence that charter schools will seriously harm the cause of public education. Some creaming may occur, but there will be too few charters to have much of an impact on the public system as a whole. Some charter school parents and students may be more satisfied, or get better educations, at little cost to non-charter schools. Letting many flowers bloom through chartering may, in the end, be the best thing that could happen to public schools. This will not be because-as charter proponents expect-charters will transform regular education. On the contrary: as charter schools face the same problems regular schools confront, they will find themselves, perhaps to their own astonishment, developing remarkably similar solutions. If this happens, the charter school armistice could evolve into something a little more like a lasting peace.