The coverage in the press gave a misleading picture of the school finance decision handed down by Judge Margot Botsford in April. The judge did not say that the state needed to throw out its current system of funding schools nor did she say that we needed to divert millions of dollars from wealthy districts to poor districts in order to remedy inequalities in school spending. Instead, she said exactly what she should have said—the education in the plaintiff districts (and, by extension, other districts across the state) is not good enough, and does not measure up to the constitutional standard (originally written by John Adams) that Massachusetts must “cherish” the education of its students.

Under the Massachusetts Constitution, it is ultimately the state’s responsibility to ensure that schools are providing a good education to their students. Toward that end, Botsford laid out, in her recommendation to the full Supreme Judicial Court, two requirements:

  • that the state determine how much schools would have to spend to meet the educational needs of all their students—particularly those at risk—and then provide the necessary funds; and

  • since money alone won’t solve the problem, that the state address the leadership and other non-monetary issues that stand in the way of quality education.

Properly understood and implemented, Judge Botsford’s decision could help us make some long overdue changes in our decade-long education reform effort, changes that could lead to major improvements in the schools.

Poor grades

The judge’s findings on school quality come as no surprise to those who follow the MCAS results. Ten years after the 1993 reform law, with the state fully meeting its foundation budget obligations and now spending some $2 billion a year more on schools than prior to the reform, only 14 percent of Hispanic and black students in poverty meet the goal of proficiency in math and only 28 percent in English (Figure 1). Proficiency rates for non-poor minorities and for poor white or Asian students—virtually identical at 26 percent in math and 46 percent in English —are also very low. These scores are up somewhat from where we started, particularly in English, and particularly for 10th-graders. We’re looking here at the percent of students who are proficient—that is, those who can function well in the modern world. This is the standard required under the federal No Child Left Behind law, and it is somewhat higher than the “needs improvement” standard required of 10th-graders for high school graduation.

The virtually unanimous assumption a decade ago was that a tripling of state education aid and the adoption of high-quality assessments would produce a sea change in educational achievement—certainly something more impressive than having only one minority student in six proficient in math and only one in three proficient in reading. It hasn’t happened.

Since the disappointing results have come primarily in educating students from minority or low-income families, it is commonly assumed that it is our big-city schools that are failing. In fact, judging by math proficiency rates (see Figure 2), suburban school districts—those with very few low-income students—don’t do much better at teaching minority or impoverished youngsters than do inner-city schools. Conversely, inner-city schools don’t do much worse than the suburbs in educating middle- and upper-class white or Asian students. (The results in English language arts tell the same story, though the rates are higher.) The difference in overall performance between suburbs and cities is not so much quality of teaching, but the fact that the suburban schools have relatively few minority students and inner-city schools have relatively few well-off students.

One of the major virtues of the No Child Left Behind federal education law is that it requires that we look at data separately for minority and low-income students. When we do so, it becomes clear that virtually all schools have failed to educate properly the students who are most at risk.

Raising the foundation

Contrary to the headlines, Judge Botsford does not call for dismantling the state’s funding system. What she does say is that one of the key elements in that system—the foundation budget—is set too low. The foundation budget is the amount (higher per pupil in high-poverty districts) that the state requires each district to spend. Botsford would require the state to determine how much money is needed to provide an adequate education, and then to appropriate it. The question is, how much too low is the foundation budget? And how much will it cost to raise it to a level sufficient to provide an adequate education to all students?

Botsford observes that districts performing at high levels spend, on average, 130 percent of their foundation budget amounts. One possible statistical inference, then, is that we’d have to raise the foundation budget by 30 percent. Still looking at the problem statistically, I’d suggest that a 10 percent increase is just as plausible. It’s true that the average spending of high-performing districts is 30 percent above the foundation. But it’s also true that the likelihood of success is no greater at 130 percent of foundation than at 110 percent.

Figure 3 compares student performance (vertical axis) with per-pupil spending (horizontal axis) for all high-poverty elementary schools in Massachusetts. Comparable value is the difference between what each student actually scored and the average score for that subject and grade of all students in the state with the same gender, poverty status, and ethnicity. Schools with positive comparable values are those where students outperform their demographic peers. Among high-poverty schools, there is a wide range of performance at every spending level. The higher performing of these schools are not necessarily the highest spending; once spending reaches about $7,000 per pupil, higher expenditure amounts do not increase the likelihood of success. The foundation budget for an elementary school with 20 percent poverty is about $7,200 per student today; it was about $6,400 in the year for which the chart is constructed, so that this $7,000 threshold is roughly 10 percent above the foundation budget.

These numbers give us a very rough idea of how much the judge’s decision could cost. The foundation budget is supported in part by the state and in part by local tax effort. Although the details of the state’s aid formula are complex, the basic idea is straightforward—for each district, the state makes up the difference between what it is required to spend (the foundation budget) and what it is expected to raise itself. This portion of state aid—the money needed to get all districts up to foundation-level spending—amounts to all but $240 million of the aid scheduled to go to school districts for the 2004­05 school year (fiscal year 2005). The remainder, which goes mainly to wealthier towns, is “extra” aid left over from when there was additional money distributed to all districts even when it was not needed to get them to foundation-budget level.

Raising the foundation budget by 10 percent would require an increase of $530 million in aid; a 30 percent increase would cost almost $2 billion more. If carried out in the 2004­05 school year, this higher adjustment would mean total state school aid of $5.04 billion, instead of the $3.18 billion currently scheduled, an overall increase of 58 percent. As the foundation budget is raised, the state’s obligation to school districts also rises, increasing required state aid. For most towns, required local support of schools is set in relation to prior year effort, new property added to the tax rolls, and overall tax wealth, and is not affected by the increase in the foundation budget. Most of the increased foundation budget, then, is the financial responsibility of the state.

At the same time, as the foundation budget increases, some of the aid that was previously “extra” becomes necessary to help even wealthier towns get to foundation. Thus, with a 10 percent increase in the foundation budget, the amount of “extra” aid statewide declines to $100 million; with foundation increased by 30 percent, only $30 million of extra aid remains. In other words, at a higher level of foundation budget, more districts with relatively high property wealth would need state aid to reach the spending requirement—which means that there is little room to finance increased aid to poorer districts by cutting aid to wealthier districts. This is illustrated in Figure 4, where the darker portion of each bar shows pure foundation aid and the lighter portion the amount of “extra” aid.

Money is not enough

The aid projections above are back-of-the-envelope estimates based on statistical observations. We can’t, in fact, know in advance how much Judge Botsford’s decision might cost to implement because her ruling placed the emphasis on improving education, not on increasing appropriations per se. Instead of a formulaic approach, she recommended to the Supreme Judicial Court requiring the state to consider carefully what communities would need to spend to meet the constitutional standard of educational excellence, and then to incorporate these findings into the foundation budget. The elements to be costed out include, in Botsford’s view, adequate funding of special education; adequate teaching of all seven state-approved curriculum frameworks to every student; adequate school facilities; and a public preschool program for 3- and 4-year-old children.

This is an imposing, and potentially costly, list of items to be recalculated or incorporated for the first time into school funding. But the judge’s insistence that the state must budget for the kind of education state law (and court rulings) require should also be viewed in light of her other key finding—that money alone won’t solve the problem. After ruling that the foundation budget is inadequate, Botsford continued: “This is not to say, however, that increases in the foundation budget alone will produce an adequate educational program in these districts. There is also a need to enhance the managerial, administrative, and leadership capacities of the districts.” This finding is consistent with the data in Figure 3 above: Once spending reaches a certain level, further increases do not increase the likelihood of success. Beyond that point, what matters is the ability to spend that money in ways that improve education.

Judge Botsford’s finding that underperforming school districts lack not just money but also the capacity to use it effectively is a very important development in the ongoing debate about educational improvement in Massachusetts. To date, the state has set higher standards, spent more money, and introduced school choice and charter schools. The comfortable assumption that these steps alone would bring dramatic improvement has not been realized; the judge’s findings underscore this. In response, however, the debate has consisted of various parties suggesting that we need more of the same—more money, or more charter schools. Paradoxically, what the debate has lacked so far is a discussion of how schools improve. The most constructive response to Botsford’s decision would be first to understand how schools improve, and only then to put together a funding and school improvement plan.

How schools improve

A good place to start the discussion of school improvement is to understand why more money doesn’t necessarily mean better education. A medical metaphor may help: Until recently, the standard treatment for severely clogged arteries was bypass surgery, which is invasive, painful, expensive, and risky. In the last few years, medical researchers have discovered that stents—particularly medicated stents—are more cost-effective ways to keep arteries open. They produce similar or better results, involve less risk to the patient, allow a speedier recovery—and cost less.

As in medicine, the key to effective school change is to use rigorous
assessment to determine the problem and research-based interventions to address it. Because of my work evaluating the Alabama Reading Initiative—perhaps the largest school improvement effort in the country, with more than 425 participating schools and thousands of teachers trained—I am most familiar with literacy instruction, but the virtuous circle of data-driven instruction really applies across all of education:

  1. Frequent assessments are made of individual student progress and of the particular deficiencies that are holding each student back.

  2. Teachers learn classroom-management skills that enable them to work with a small group of students while other students are constructively engaged in independent or small-group activities.

  3. Teachers use research-based interventions to address the deficiencies they identified in step 1, and are able to do so in the small group settings they learned to manage in step 2.

  4. Progress of struggling students is monitored frequently—weekly, for those farthest behind—and this monitoring is used to determine whether the interventions in step 3 are working. If not, the teacher needs to try something else and—with help from coaches knowledgeable in the subject—look for new strategies.

There are four other elements that contribute to the success of this kind of program: support and inspiration from the principal; leadership roles for teachers in monitoring student progress, devising new curricula, and helping each other implement change; consistently high expectations from the faculty and principal as to what their students can accomplish; and support from outside coaches who lay out a vision of change and insist that the school stick with it.

The other day I received a letter from the intermediate school (Grades 4­6) in Coosada, Ala. In response to the in-depth training they received this fall, teachers held a series of discussions on what was working in their school and what was not. They’ve totally revamped their reading program, adopted a new school schedule for next fall, and doubled the time available for interventions with struggling readers. They’ve also begun a shared teaching program—two teachers from different grades and/or subjects teaching a class together—breaking down the isolation in which most teachers work. This is the kind of dramatic change we need to see here in Massachusetts.

In rural Pine Apple, I walked through the local school (all black, all poor) with the principal. As we entered each classroom, I asked him to point out the struggling readers—and he knew exactly who they were. He meets with his teachers twice a month to discuss the strategies they’re using with each one, and he couldn’t wait to share with me the progress-monitoring results that showed the gains they’re making.

Teachers are buying into data-driven instruction.

We don’t need to go to Alabama to see how this can work. As part of the federally funded Reading First program, Massachusetts requires participating schools to administer a reading assessment to all students at least three times a year. At an inner-city school south of Boston, the Reading First coach used these test results to show teachers that 75 percent of their students couldn’t decode —that is, they couldn’t automatically and fluently translate symbols on the page not just into sounds but also into words, sentences, and paragraphs.

I’m sure those teachers already knew that most of their students were struggling. But they hadn’t previously known that decoding was the specific problem. If you don’t know the problem, you can’t fix it. Particularly in literacy, there’s extensive research on the kinds of interventions to use for most of the difficulties teachers encounter. The coach helped the school pick out and purchase intervention materials specifically designed for decoding prob-lems and helped the teachers learn to use them. As they begin to see their progress-monitoring results improve, teachers are buying into the whole notion of data-driven, research-based instruction.

Chicopee schools are moving aggressively to improve reading instruction. Outside coaches are showing teachers where their students’ specific needs are. They are purchasing specific intervention materials tailored to the kinds of deficiencies they’re finding. Many of these are 30-minute, highly structured lessons prepared by leading researchers in the field and designed to work on a specific skill. Title I teachers, special education teachers, and paraprofessionals are trained to give these lessons to the students who need them. For example, a key remedy for fluency difficulties is to have a passage read aloud by a fluent reader, then to have the student re-read the passage—perhaps several times. This can be done by paraprofessionals or adult volunteers who get a little training. As teachers see their students progress, they get interested and want to learn more.

The principal at Chicopee’s Bow School has put together a kind of primer, documenting some of the most frequent reading deficiencies, as identified by the reading assessment, along with a set of suggestions for how to address them. One of the key features here is that results of these assessments are instantly available to teachers, so they can immediately see progress and begin to use assessments to measure whether their interventions are working. This empowers teachers and gives them a tool they’ve never had before.

It’s impossible to overstate how far behind children from non-literate homes are when they come to school. By one account, the typical kindergartener from suburbia knows four times as many words as her inner-city counterpart. So minority and low-income students will need extra time to work on vocabulary, language structure, and grammar if they are to have any hope of reading with comprehension—or of mastering science, math, or history. But the Chicopee principal has the right idea—assess progress, identify student deficiencies, and offer appropriate interventions.

It’s commonplace in some circles to argue that unions present a critical obstacle to success. That is not my experience, at least with the kind of program I’ve suggested here. To be sure, teachers expect to be paid for the extra time they put in, but this is not unreasonable. As long as their extra effort is recognized, they are generally happy to participate and not to watch the clock. In schools I’ve visited in Alabama and in Pennsylvania, I’ve found local union leaders who have been the strongest proponents of this kind of change.

Prescription for improvement

If the SJC accepts Judge Botsford’s recommendations, the high court will order the state to calculate the cost of providing an adequate education, and also to propose “measuresŠthat will provide meaningful improvement in the capacity of these local districts” to provide that education. I suggest that we put aside for a few months discussion of how much more we need to spend or how to restructure the Department of Education. Instead, let’s start by developing a shared vision of what it takes to turn around schools in large numbers. Once we do this, the funding and the structure will follow logically.

Everyone knows that you can turn around almost any school by hiring an outstanding principal. But, except in Lake Wobegon, you can’t insist that all principals be above average! Any plan for improving the educational outcomes of large numbers of poor children has to be able to work in any school, by engaging the interest and enhancing the skills of typical principals and teachers.

The power of the assessment/small group/intervention model—as our Department of Education is beginning to apply in Reading First and as my clients in Alabama are using—is that it has already turned around dozens of schools. To accomplish turnaround on a wide scale, we need to practice prescriptive intervention—that is, the department has to have a well-conceived model of school change, and has to insist that participating schools follow that model.

This is not how we’ve typically done business in education. Until now, we’ve believed in local autonomy, with individual schools and school districts free to conduct their programs with as little guidance as possible from the state’s Department of Education. But teachers and principals don’t know what they don’t know. If they didn’t know their students couldn’t decode, and didn’t know the research on decoding interventions, and didn’t know about assessments and small-group instruction, they couldn’t improve reading results—even with more money.

The great majority of teachers and principals went into education because they want to help kids. They may grumble at first about having to follow a prescription they didn’t write, but when they see it work they’ll become strong supporters. That’s what happened in Alabama. The ARI just received $40 million in state money for next year, up from only $12 million—this in a state that’s practically broke. It wouldn’t have happened without overwhelming support from teachers, who are thrilled with the improvement they’re seeing.

Back to money

Once we get behind this kind of prescriptive model, the implications for funding will be readily apparent. The Department of Education will need additional funds for outside coaches, development of instructional materials, and teacher training academies. Schools will need money for assessments, teacher training, common planning time, structured intervention materials and instructional programs, and school-based coaches.

How much will it cost? A detailed answer should follow, not precede, adoption of a change model. But a very crude calculation shows that the model I’ve suggested here would cost something like $200,000 for a school of 400 students—or roughly $500 million statewide. To work properly, this kind of program requires reasonably small classes, particularly in the primary grades in high-poverty schools, and to that end we’d have to restore perhaps $50 million to $100 million of cuts made in recent years. In total, this is roughly equal to a 10 percent increase in the foundation budget. In addition to this boost in operating budgets, it’s clear from Judge Botsford’s ruling that the state will need to address overcrowding and lack of adequate facilities in some schools across the state. Both the governor and the legislative leadership have made proposals to accelerate school building construction; as this is written, these proposals are being debated in the Legislature.

It’s not clear whether $600 million or $700 million spent on structured intervention and reducing class sizes in high-poverty schools, as I’ve laid out here, would satisfy Judge Botsford’s requirement that schools work toward excellence in all of the subject areas encompassed by the state’s curriculum frameworks. But anyone who knows schools understands that there’s only so much you can change at any given time. A program of comprehensive school change, based on the virtuous circle of assessment, small-group instruction, research-based interventions, and feedback from assessment to instruction, will fully absorb the energies of faculty and principal alike and, even in the most capable of schools, will take three or four years to accomplish.

Since reading is central to all other learning, such an effort should initially be focused on reading. (One of the most interesting of my findings from the Alabama Reading Initiative is that by fifth grade, ARI schools outperform nonparticipating schools not only in reading but also in math.) When schools are successful at teaching all students to read, fewer students are referred to special education and many difficult behaviors disappear (older students who can’t read often act out to hide the fact that they can’t follow the lesson).

In light of all this, it would make sense for the state to make an interim response to the Botsford order—a response that says we’ll spend three or four years focusing on data-driven instruction, primarily in literacy. At the end of that time, teachers will be ready to move on to other areas, and we can determine whether schools are then able to meet constitutional standards across the board or whether additional resources and training are needed. Paradoxically, we’ll make more progress in the long run with an incremental approach paced to the realities of today’s students and to the difficulties of the task we’re undertaking.

There is no doubt in my mind that prescriptive intervention programs can work. It takes longer than we’d like, because kids from disadvantaged homes start so far behind. Along the way, we’ll uncover new problems and have to figure out how to address them. But we’ll also generate a wave of enthusiasm and energy from teachers who will be thrilled to see their kids moving forward.

Edward Moscovitch is principal of Cape Ann Economics.