Are inadequate teachers causing schools to underperform? A 1998 Harris poll finds that, by a ratio of over six to one, the public thinks improving teacher quality, not instituting vouchers or charter schools, is the key to school reform, more important than standardized testing, better curricula, or smaller classes. President Clinton complains that there ought to be a "process for removing teachers who aren't competent" and proposes denying federal funds to schools that hire unqualified teachers. We "can no longer fiddle around the edges of how we recruit, prepare, retain and reward America's teachers," says Education Secretary Richard W. Riley. "A growing number of school districts are throwing a warm body into a classroom, closing the door, and hoping for the best."
We have bemoaned teacher quality at least since "A Nation at Risk," the 1983 report of the Reagan administration's Department of Edu cation, concluded that America was economically uncompetitive with Europe and Japan because of inferior schools. That charge soon proved ridiculous. And the report's claim that "teachers are being drawn from the bottom quarter of graduating high school" students was always nonsense; virtually no youngsters from the bottom quarter of high school seniors go to college, much less become teachers. Yet the prescriptions of "A Nation at Risk" have, if anything, gained in influence since the report was published—particularly the insistence that our poorly trained teacher force desperately needs upgrading.
Today a widely accepted syllogism goes something like this:
- Student performance is unsatisfactory because most teachers are themselves pretty dumb. (If they're not the least talented high school grads, then they must be the least able college students.)
- Traditional certification encourages hiring academically inferior teachers because it requires courses in pedagogy, not academic content.
- Once inferior teachers are hired, tenure and unions make getting rid of them impossible.
- Thus, to improve student outcomes, districts need to break the lock of both education schools and unions, by recruiting bright college grads uncorrupted by "methods" courses and by allowing principals to fire teachers who can't raise student test scores.
Versions of this story are now being advanced by both the New Democrats' Progressive Policy Institute (PPI) and the Thomas B. Fordham Foundation, a conservative group led by former Reagan education official Chester Finn. Their get-different-teachers approach to education reform seems to have enormous appeal, perhaps because most parents remember at least one horrible teacher who may have had no business in the profession. But we don't really know if children typically must endure one bad teacher—or, say, five—out of the 15 or more they usually have before high school. If it's only one, quality control in education would seem to be on a par with well-run institutions in other fields. In this case, what's needed is not a whole new teaching corps but ongoing efforts to improve overall instructional quality and to weed out the few teachers who don't belong. If, on the other hand, children typically have more than an occasional bad teacher, gross dissatisfaction with public schools may be justified. This is precisely what the critics assume, but there is no basis for their assumption. Anecdotal and polling evidence actually suggest the opposite: In the Harris poll, most Americans (69 percent) considered their communities' teachers to be well or even highly qualified. In the absence of data, it is quite premature to suppose that the problem of teacher quality is as dire as the critics say—or as simple to solve.
In fact, nobody has figured out how to disentangle the effects of teachers on school achievement from the effects of families, economic circumstances, or school culture. [See Richard Rothstein, "Charter Conundrum," TAP, July-August 1998.] Even if we could isolate an individual teacher's responsibility for student performance, we couldn't quantify precisely what good teachers do to generate achievement and thus couldn't easily test anyone for it. Elevating overall teacher effectiveness and removing poor teachers from the profession require one-on-one evaluations of teachers and teacher candidates by qualified professionals. This kind of nuanced supervision is a far more expensive approach than reformers like Fordham and PPI are prepared to contemplate.
A pernicious mix of myth and reality underlies the campaign against teacher inadequacy. It needs to be examined.
Bottom of Their Class
Consider the now-conventional assertion that, as PPI puts it, "students who enroll in teacher education programs in U.S. colleges tend to have lower scores on SAT and ACT exams than those in virtually all other programs of study." This claim usually originates with analyses of voluntary questionnaires distributed by the Educational Testing Service (ETS) with SATs, showing that high school juniors and seniors who plan to major in education have lower average scores than classmates who plan arts and sciences majors. Yet high schoolers announcing intended majors may eventually major in other fields; majors in education may not attempt to become teachers or may fail qualification tests if they do try to teach; and arts and sciences majors may teach after taking courses in pedagogy or (increasingly) after earning a master's degree in education. High schoolers' SAT questionnaires provide no useful information about the abilities of actual teachers. Another common analysis—of the SAT scores of entering freshmen at colleges that graduate teachers—isn't much help either.
ETS studies of actual teachers in fact show their academic abilities comparing favorably with those of other professionals. Analyzing the 1992 National Adult Literacy Survey, ETS found that the verbal scores of teachers are higher than those of managers but similar to those of lawyers, engineers, accountants, and social workers. In quantitative skills, teachers are on a par with most other managers and professionals. ETS also looked at the high school SAT scores of those who later passed Praxis tests, used by many states for licensing teachers. High school teachers who passed Praxis tests in math, social studies, foreign languages, science, and English had higher SAT scores than college graduates generally. Teachers who passed Praxis tests in physical education, special education, elementary education, and art and music had lower SAT scores.
Though claims of low teacher ability are unfounded, bright teachers with higher than average scores do generate higher student achievement. We know, for instance, that students with high math and reading scores are likely to have had teachers who themselves had high scores. James Coleman's 1966 analysis "Equality of Educational Opportunity" made this argument. It was reinforced in a widely circulated study of Texas student achievement, published by Kennedy School scholar Ronald Ferguson in 1991. Reports by Dominic Brewer of the RAND Corporation and his colleagues confirm that teachers with high test scores probably improve their students' scores.
But it doesn't follow that good teachers succeed because they have a deeper understanding of the subject matter or that they can succeed without the pedagogical skills that teacher-training institutions should impart. There's a plausibility to the notion that you can't teach what you don't know, but most teachers cover material far less advanced than that distinguishing high from low SAT scorers. It's just as plausible that brighter teachers succeed not only because they have superior content knowledge but also because they can better diagnose learning difficulties; are more flexible in adapting to students' varied learning styles; make more balanced decisions about when to work with students individually, in groups, or in whole classes; and can judge when it makes sense to wait patiently for a child's answer and when to move on. College grads with high SAT scores but no training in pedagogy are unlikely to have these skills.
Bypassing the Ed Schools
There may nonetheless be merit in bypassing traditional undergraduate education schools. In 1986 the Carnegie Corporation's Task Force on Teaching as a Profession recommended that all teachers major in academic subjects as undergraduates, not in education, and then obtain a master's degree in pedagogy before being licensed to teach. A consortium of university education deans (the Holmes Group) advanced a similar proposal, that instructors with a bachelor's degree work only under the direct supervision of professional teachers who have postgraduate pedagogical training. In 1996 a National Commission on Teaching and America's Future, organized by Linda Darling-Hammond, revived and promoted these recommendations. They still make sense.
Recruiting people who didn't originally train to be teachers also makes sense, where programs include serious pre- employment pedagogical training and intense supervision of initial teaching. However, many alternative-certification programs offer neither. Best known is Teach for America (TFA), founded by Princeton graduate Wendy Kopp, who took pride in her lack of professional training and teaching experience, arguing that states should "abolish licensure laws altogether." TFA assumes that enthusiastic college graduates can be effective after attendance at its summer workshop, mostly designed and led by other recent recruits. Indeed, some TFA teachers have been successful, but others have not. Faced with the realities of classroom life, TFA recruits bail out more rapidly than typical beginning teachers. Many TFAers graduate from elite universities prior to being assigned to inner-city schools. As Linda Darling-Hammond observes, "Never having had difficulty learning themselves, they cannot figure out . . . what to do when a lesson does not succeed." TFA's defenders argue that, given the shortage of inner-city teachers, the alternative to a TFAer is not a properly trained teacher, but a substitute with no training at all, and that some TFA recruits remain in the profession, continue to acquire training, and become unusually creative teachers. But despite these successes, long-term teaching stints have not been the program's primary goal. Instead, TFA has promised idealistic college grads meaningful teaching experiences for a couple of years before they embark upon "real" careers. While senior teachers are not always better than junior ones, and teachers do not gain effectiveness in direct proportion to experience, skill does tend to grow during a teacher's first 7 to 10 years. If teachers generally don't hit their strides for nearly a decade, then TFA is not the best model.
Alternative certification for older career-changers also should require rigorous professional training. Experience as an accountant or business manager can't substitute for knowledge of child psychology, pedagogical theory, and classroom management tools. One promising approach was the federal Troops to Teachers program implemented in 1994. Troops to Teachers provided $5,000 stipends to military personnel with a bachelor's degree to enroll in teacher-training programs, and then it subsidized districts to hire those who gained certification. While program quality varied with the rigor of state certification programs, there was less attrition among the 3,000 Troops to Teachers than among TFA recruits. The ex-military personnel were also more likely than existing teachers to have hard-to-staff specialties like math, science, and special education; to be members of minority groups in which teacher shortages are especially acute; and to be willing to teach in inner cities or rural communities. The Clinton administration has proposed expanding this program, but its future in Congress is uncertain.
Mentoring on the Cheap
Even with pedagogical training, talented recruits flounder without support. Critics and policy makers alike agree on the need for a more highly paid corps of "mentor teachers" to develop young teachers' competency, evaluate more senior teachers, partner with less experienced teachers to remedy deficiencies, and recommend additional training or, in extreme cases, removal from the profession. What is not widely acknowledged is that this reform would create a new and expensive school supervisory structure. Because mentors cannot observe, diagnose, and collaborate with junior teachers while instructing their own students, the costs of mentoring include not only bonuses paid to mentors, but also salaries for replacements in their classrooms.
Curiously, many mentoring advocates also denounce what they regard as schools' excessive administrative spending. Yet schoolteachers are the most undersupervised professionals in the nation. A recent law school graduate, hired as an "associate," typically works for seven years preparing briefs, meeting with clients, even trying cases under the tutelage of senior partners. Partners rarely (except in some New York megafirms) oversee more than one associate at a time, and they may spend 25 percent of their time on this mentoring. Similar relationships characterize accounting, architecture, and medicine. Few corporate managers have more than four or five "direct reports" (the junior executives under their supervision). At The American Prospect, five senior editors devote much of their time to mentoring four writing fellows plus other junior editorial employees—about a 1:3 ratio.
But public school supervisory ratios are typically closer to 1:20. A principal may supervise 20 teachers or more, evaluating each by making a few brief classroom visits. In larger, better funded schools, assistant principals may help, but the ratios remain much the same. Training and evaluation compete for administrators' time with disciplinary crises, parent meetings, and administrative tasks. Teachers who claim they're not "treated like professionals" are right. No junior lawyer would ever handle 25 important client cases behind closed doors, with only twice-yearly peeks by supervising attorneys.
Yet imagine a system in which skilled teachers are mentors to junior colleagues. Imagine that each mentor could, if functioning full time in this role, simultaneously supervise five junior teachers by demonstrating model lessons, observing junior teachers at work, jointly diagnosing student needs, recommending and discussing professional literature. Imagine that, when removal of junior teachers is the only reasonable option, mentors are given time to document deficiencies, prepare necessary paperwork for due-process hearings, and either counsel junior teachers to resign or testify regarding their unsuitability. Imagine that half the regular teachers in a school are sufficiently senior and talented not to require such supervision and that mentoring roles continue to be shared with principals and assistant principals. Still, this model could require the number of school professionals to grow by 10 percent. At current salaries—accounting neither for the next decade's anticipated enrollment growth nor for efforts to reduce class size—the annual cost of replacing mentors in their classrooms could be more than $13 billion.
Mentoring is the one reform likely both to improve average teacher quality and to remove the worst cases of malpractice. But nowhere in America have enough mentors been relieved of teaching responsibilities so they may do either job effectively.
To be sure, incompetent teachers are hard to remove [see Kate Cambor, "Bad Apples," TAP, September-October 1999], and it's not just union protections and public employee codes that are to blame. In a recent study of inner-city Catholic schools, Martin Carnoy, Louis Benvenista, and I were surprised to discover that, even here, underperforming teachers were rarely dismissed because archdiocesan legal departments, fearing lawsuits, imposed due-process protections nearly as extensive as in the public sector.
Horror stories abound of districts spending small fortunes on hearings to dismiss incompetent teachers. But districts' hesitancy to fire incompetent teachers, when counseling fails, has little to do with the costs of arbitration. In fact, if school districts effect one involuntary termination per thousand teachers per year, and if formal procedures cost $100,000 in legal fees (far less is typical), this cost is only $100 per teacher per year—worth every cent if it gives greater credibility to more routine mentoring procedures.
No, it's the costs of effective supervision that frustrate good personnel management. Most discharge-for-incompetence cases fail because districts are forced to defend inconsistent practice. Typically, a new tough principal takes action against a teacher whose incompetence was tolerated for years. Or a principal attempts to remove a teacher whose incompetence is no greater than that of others in the district. Winning arbitrations requires demonstration of consistent practice, which is impossible unless every teacher is supervised and is given opportunities to improve—or progressive warnings about failure to do so. None of this can happen with current school supervisory ratios.
Some districts (Rochester, Toledo, and Columbus are best known) have implemented peer-review programs to enhance supervision, but under most state bargaining laws, peer-review mentors cannot recommend discharge of incompetent teachers without themselves becoming supervisors, legally ineligible for union membership. Consequently, mentors have not been given this essential power, and peer-review districts have had little success in removing incompetent teachers. The lack of results threatens to discredit the whole idea of mentoring.
Overwhelmingly, however, it is the looming shortage of teachers that defeats virtually every reform idea for improvement of teacher quality. In the next decade, the baby boom "echo" will add 1.5 million schoolchildren to the nation's classrooms. At the same time, many existing teachers will be retiring—one-fourth of today's teachers are over age 50, two-thirds are over 40. These trends alone require hiring 200,000 new teachers annually over the next 10 years—five times the recent rate.
Moves to reduce class size will further exacerbate this shortage. Based on limited but probably valid research showing that students benefit from small classes in primary grades, the states and the Clinton administration are rushing to fund this reform. In the context of the current shortage, this could be disastrous, as California recently demonstrated.
In 1996—after years of stagnant school spending, with classes among the largest in the nation—California suddenly appropriated $1.5 billion to reduce class sizes to 20 in grades K-3, with additional appropriations in subsequent years. But when classes were suddenly reduced from 30 to 20, demand for primary grade teachers instantly jumped by 50 percent. With rapidly rising enrollments and too few qualified teachers, the new policy caused a calamitous churning and redistribution of teachers.
Vacancies appeared everywhere. The best and most qualified teachers could choose districts in which to work, and they took newly created positions where conditions were considered more favorable—fewer children "at risk" of failure, more coming to school ready to learn, fewer discipline problems—or in districts closer to their suburban homes. As the best teachers abandoned inner-city assignments, urban schools had to hire teachers without credentials or training. Statewide, one-fourth—and in Los Angeles, two-thirds—of new teachers hired to reduce class sizes lacked credentials. The most disadvantaged children traded large classes with more skilled teachers for small classes with those less skilled.
Wishful Thinking about Salaries
Today's reformers are much taken with economists Dale Ballou and Michael Podgursky, who argued in a 1997 report that widespread teacher inadequacy can't be remedied with raised salaries. Union-inspired salary schedules, they said, will prevent schools from focusing new funds to attract more effective teachers. Rather, districts will distribute raises among all teachers on the schedule, wasting money on low-ability teachers already employed. Ballou and Podgursky claim that raising salaries could even result in lowering teachers' average effectiveness because higher pay would dissuade existing teachers from quitting; seeing fewer vacancies, more talented college students would decide against pursuing teacher training.
The warning about too few vacancies to attract high- ability candidates is curious, given the gargantuan teacher shortage we are facing. In such a market, teacher salaries cannot help but be a crucial factor in recruitment. Because professions are not fully comparable, we can't estimate precisely how much teacher pay must increase to influence both career changes and the initial decisions of bright students contemplating alternative careers. But managers and administrators, for example, whose average literacy level is below that of teachers, currently earn, on average, about $30,000 more per year.
It will also take substantially higher salaries to address the qualifications gap between suburban and inner-city teachers, accelerated by union contract provisions that give teachers a seniority right to fill vacancies in more "desirable" schools. Thus, within metropolitan districts, experienced teachers flock from inner-city, poor, and minority schools to predominantly suburban, middle-class, and white schools. Making matters worse, cascading seniority moves—each opening a vacancy for another teacher to occupy—may not conclude until the beginning of the academic year, leaving districts to fill last-minute vacancies with uncredentialed applicants in locations where teaching skill is most critical.
It is no use to demand, however, as many critics do, that unions abandon these procedures. Teachers who have accumulated experience should have opportunities, if they wish, to teach where discipline is easier, children come to school more ready to learn, and schools (and teachers) get public credit for students' easier-to-achieve successes. Not all experienced teachers want this, but those who do will only seek openings elsewhere if they are prevented from exercising seniority rights to choose assignments within their districts, or they will abandon the profession altogether.
The alternative is pay differentials for hard-to-staff schools. Some local unions have agreed to these, although amounts have usually been token, insufficient to equalize teacher experience between schools attended by poor and middle-class children. New York City, however, has recently offered 15-percent bonuses to teachers who volunteer for its 40 lowest-scoring schools. For experienced teachers, this could amount to $7,500 a year. If substantial numbers of teachers accept the offer, and if it is extended to all inner-city schools now staffed by inexperienced teachers, massive new funds will be required.
In fields with the most serious teacher shortages—math, science, bilingual studies, and special education—the solution is the same. There is even stronger sentiment in public education against these kinds of pay differentials. Unions earlier fought to establish single-salary schedules to replace the separate scales for better-paid (mostly male) high school teachers, and lower-paid (mostly female) elementary teachers. But today, paying high school physics teachers more than fourth-grade classroom teachers does not imply that high school physics is more important than fourth-grade reading. Rather, it signals a market imbalance requiring a market response. If college students and other professionals respond to higher pay by flocking to hard-to-staff schools and fields, differentials can always be reduced.
PPI and Fordham believe that teacher quality can be ensured only if principals and teachers are held accountable for student performance—rewarded for high achievement and replaced when performance is inadequate. They are right. But how to measure performance? The most popular methods— standardized reading and math tests—are dangerously flawed. Because we want a variety of outcomes from schools (good citizenship, ethical behavior, creativity, health, physical fitness, teamwork, and personal responsibility as well as literacy and numeracy), holding principals and their teachers accountable only for math and reading skills, which are the easiest skills to measure, creates perverse incentives to ignore the rest.
Even if we look only at basic skills, it's nearly impossible to judge a school staff's effectiveness by test scores. Most score variation results from family influences, not school effects. A principal and teachers in a middle-class community with high test scores may be less effective than a principal and teachers in a disadvantaged community with low scores. In poor communities where students move frequently, it's hard to know which school is responsible for which students' outcomes. Scores in any year are the cumulative product not only of that year's teachers but also of those in each prior grade.
For 75 years, state legislatures have made sporadic attempts to implement performance bonus plans for high-performing schools and merit pay plans for superior teachers. Most have been abandoned after short trials, defeated by controversy over whether award criteria were really within schools' or teachers' control. The conservative Fordham Foundation along with the Education Trust, a liberal advocacy group for poor and minority children, now argues that districts can overcome these problems by utilizing statistical methods developed by Professor William Sanders of the University of Tennessee. Sanders compares teacher (and school) effectiveness by measuring the "value added" by each teacher each year, the gain between last year's and this year's scores. Yet "high value-added" Tennessee schools seem more often to have high proportions of middle-income students. Perhaps Sanders is onto something, but the possibility remains that his method measures student social class as much as teacher or school effectiveness.
We have little choice, therefore, but to rely on supervisory judgments regarding which teachers are effective and which should be removed, and on superintendents' judgments regarding which principals are suited to make such decisions. Many superintendents have little understanding about how to select and evaluate good principals, arguably their most important task. This problem must be solved before teacher quality can be addressed successfully.
Holding schools accountable for results requires more than frequent testing. It requires adequate supervision, consistent leadership, and well- designed incentives. It requires schools organized around a nuanced conception of high achievement, with principals and mentor teachers chosen because of their ability to establish a school culture that responds to rewards for high achievement, unquantifiable though it may be.
For this model, we've got to fund salary increases to attract college students to teaching, though some of this new money will unavoidably "leak" to teachers already employed. Money is needed for higher salaries in hard-to-staff fields and locations, recruitment of even more teachers to replace mentors who move out of full-time classroom roles and into supervisory ones, class-size reduction for the most disadvantaged children, and expansion of public schooling to preschool years, especially for at-risk youth. Even if average teacher salaries increased by only $10,000 annually to compete with other professions, the bill for these reforms could easily total $50 billion a year, politically a nearly inconceivable increase of almost 20 percent in what schools presently spend. Attacks on teacher ability are cheap; providing the teachers our schools need will not be.