CHRISTOPHER T. CROSS
This conference brings together statisticians, researchers and policy makers, three groups that must talk together, but—as we know—sometimes have difficulty finding the time.
But this type of communication is vital. Surveys and other data collection systems must provide the information policy makers need at the time they need it and in a form they can use. For their part, policy makers need to be more clear about what their needs are and about what information is most useful. This conference provides an opportunity to engage in this sort of give and take.
The subject of this conference—teacher supply, demand, and quality—is clearly of central importance to education. The implications of the relevant statistics and research affect the most important determinant of learning—classroom teachers.
Yet this is a field still very much in its infancy. As seen in the papers presented at this conference, debate continues over such basic issues as what to measure, how to measure it, and even when to measure it. Even more work remains to be done in designing models that predict shortages ahead of time, so that policy makers can respond in time.
Nonetheless, progress is being made on a number of important issues. We have learned, for example, that yearly attrition rates can be deceptive for predicting future needs since many departing teachers eventually return to teaching after a few years. Nearly 40 percent of current teachers have taken just such a break and returned to the classroom (Feistritzer, 1990).
We have also learned that, while aggregate data is important, more attention must be focused on key subject fields that may face particular
difficulties. We know, for instance, that two-thirds of secondary school principals are having a difficult time hiring chemistry and physics teachers. Other studies have shown that many teachers in these subjects lack a degree in the field.
One policy tool several states are experimenting with for dealing with such shortages is alternative certification for teachers and principals. Such alternative routes can bring to education talented individuals who, while they may lack formal certification, may nevertheless be well qualified. This is especially important in critical shortage areas such as science, mathematics, and foreign languages.
But better, more systematic data on alternative certification (and its potential as a tool for bringing needed teachers into the classroom) are needed. Policy makers look to educational statisticians and researchers to provide the information needed to weigh such matters.
Action to ensure an adequate supply of qualified teachers is also under way at the federal level. One example is the Federal Coordinating Council for Science, Engineering, and Technology (FCCSET). Through its Committee on Education and Human Resources, FCCSET seeks to coordinate the myriad science and education programs of several federal agencies and departments into a coherent national strategy. FCCSET has made teacher education and training its number one priority.
This leads to a subject that is critical—teacher quality. Teacher supply and demand are only part of the picture. Even if we have enough teachers in critical subjects, even if teachers know the subjects well that they are teaching (which we know is not always the case), subject knowledge does not necessarily translate into teaching quality. We need indicators that give us a true picture of teacher quality, which includes teaching practices.
Studies tell us that today's teachers still tend to rely on many of the same methods teachers used at the start of this century. They "deliver" education through lectures, and they ask students to repeat lessons back on tests, often simple multiple-choice tests.
This common practice reflects an inadequate understanding of the nature of knowledge and results in a flawed approach to learning. We now know that to attain true understanding, students must be able to use facts to think and analyze for themselves, not just to recall isolated bits of data. Such new, and more demanding, expectations require approaches to teaching that are more interactive—approaches that engage students in thinking as a matter of course. But educators are expected to lead a charge into territory many of them have never experienced.
As we move toward what Lauren Resnick calls "the thinking curriculum," we face a dilemma. Teachers were taught according to the old ways we are trying to leave behind. And they have internalized the premises underlying these practices into deeply held attitudes and beliefs. Teaching
practices are drawing lots of attention lately. They should. Changing them is vital to our objectives for improving education.
During the 1980s, we decided, as a nation, that all children can learn. And that all children can learn to think. That was a main point of the national education goals. As President Bush and the governors agreed, "All our people, not just a few, must learn to think for a living." Never before has a nation set its sights so high for all its people—and all its children.
States are struggling to implement these very basic changes. California has found that one of the greatest obstacles to implementing the impressive curriculum embodied in its Frameworks and other reform efforts has been the difficulty teachers experience in revising their assumptions about how children learn. As they begin to make this shift—through training, professionalization, etc.—statistics derived from surveys can help in the following ways:
They can help us see how teachers adapt to, implement, and enact changes;
They can allow us to gauge teacher attitudes and practices;
They can establish a baseline;
They can help us monitor changes in those attitudes and practices over time; and
They can help us examine these changes in relation to shifts in policy.
As our changing understanding of learning demonstrates, teachers and other educators need to become consumers of education research. They must stay in touch with the research community so that they can implement, or at least experiment with, new findings such as "wait times" and new approaches to student assessment.
Much remains to be done to bring this about, both on the part of teachers and on the part of those producing the research. It might be helpful to consider research consumption as a component of teacher quality. How aware are teachers of relevant research and how much are they experimenting with new approaches in their classrooms?
One final note. In her paper, Mary Kennedy makes a recommendation that is particularly worth considering. She recommends that National Assessment of Educational Progress (NAEP) surveys include context-specific questions that tap teachers' basic educational beliefs and values. Such questions, she points out, "could improve our understanding of teachers' fundamental orientations toward teaching and the kinds of outcomes they promote in their students." If tied to NAEP, responses to such questions will provide a means for examining the link between teacher attitudes and student performance. Student learning, after all, is the critical feature of education—the one we all care about most.
It is the potential for providing policy makers with useful, actionable data that we, in the Department of Education, see statistics and research on teacher supply, demand, and quality—and the proceedings of this conference—as so important.
Feistritzer. C. Emily 1990 Profile of Teachers in the United States, 1990. A National Center for Education Information survey. Washington, D.C.: Feistritzer Publications.
EMERSON J. ELLIOTT
The National Center for Education Statistics (NCES) was pleased to sponsor the Conference on Teacher Supply, Demand, and Quality (TSDQ) and this collection of papers based on the proceedings. This publication represents a continuation of work undertaken by the National Research Council that focused on the supply, demand, and quality of precollege science and mathematics teachers. The purposes of the conference were to reassess the status of TSDQ a year after the 1990 report Precollege Science and Mathematics Teachers. Monitoring Supply, Demand, and Quality to update our understanding, and to help establish an agenda for future work in this important facet of public education.
Several important recommendations for NCES were forthcoming during the course of the Research Council's earlier work that we have accepted and implemented. For example, it was recommended that NCES should suspend projections of teacher supply until better data were available and better models were built. Accordingly, we did suspend such projections. Other recommendations we acted on pertained to collection of data on recent college graduates, schools, and the staffing of schools. We look forward to receiving further useful and compelling recommendations as an outgrowth of the conference.
Teacher supply, demand, and quality are of particular importance to the mission of NCES because they are built into our authorizing legislation. Back in 1974, NCES was charged with conducting surveys to determine the demand for and the availability of qualified teachers and administrative personnel, especially in critical areas. This charge was enacted about the time Washington concluded that there was a teacher glut, and that the Edu-
cation Professions Development Act should be abolished because it was no longer needed.
In 1986, the interest in teacher supply, demand, and quality—especially for mathematics and science teachers—was renewed. Congress gave NCES the additional assignment to address needs for teachers in various subject areas; to examine possible teacher shortages; to identify the sources of teacher supply; to monitor the teaching force in terms of its demographic characteristics, academic qualifications, job preparation, experience, and skills; and to determine the rate at which teachers leave the profession along with their reasons for leaving.
At the present time, much attention and energy are being devoted to the development and measurement of progress toward national education goals. The performance of U.S. students in science and mathematics is one topic of great concern, and this relates to the supply and quality of science and mathematics teachers. How to measure progress toward achieving national goals and the quality of teachers are still unresolved issues.
The NCES Perspective
NCES hopes this conference will produce three benefits for its work. First, we expect it will provide us with an opportunity to keep abreast of public policy and methodological issues in TSDQ. Second, it should provide us with an opportunity to continue dialogue about TSDQ among policy researchers, educators, and those in federal and state statistical agencies. Finally, it will provide an opportunity to take stock of innovations recently introduced into NCES programs. We are interested in your assessment of them in terms of strengths and limitations, and whether they measure up to your expectations. Perhaps more important, we hope to receive your advice about how our existing programs and surveys should evolve in the future, about new components or surveys that we should consider adding, and about how value might be added to our data bases by integrating or linking them.
The scope of our interest is typified by the structure of the conference. It is organized into discussions first on TSDQ policy issues, then on modeling teacher supply and demand behavior, next on the utility of national data bases for supplying the data needed by the models, and last, on state teacher data bases.
In terms of policy issues, it is interesting to note that, in the last few years, there has been less concern about teacher shortages; rather, the major policy issue is now becoming teacher quality. The Kennedy paper does a very good job of elucidating the quality issue. While we recognize the importance of teacher quality, we also recognize that our metrics for measuring teacher quality are woefully inadequate at this time. Your guidance on how we might proceed in this area is sought.
With respect to modeling teacher supply and demand, the Barro paper
was quite useful in summarizing the progress that has been made since he wrote a similar paper several years ago for the first National Research Council report on teacher supply and demand. NCES would value your advice on what our role, along with that of the Office of Research in the Office of Educational Research and Improvement and the National Science Foundation, ought to be in stimulating (and perhaps even funding) research into developing better TSDQ models. We at NCES clearly have a stake in the further development of models relevant to our teacher supply and demand projection methodology. In addition, we might stimulate TSDQ research with state data bases. This would be within the scope of our mandate in the National Cooperative Education Statistics System, wherein we have authority to undertake research and demonstration projects.
Although NCES has a number of data bases relevant to TSDQ, they are managed in different parts of the organization. It is therefore very useful to have someone scan across and analyze these data bases, including other existing national data sets, to give us a better perspective of how they all contribute to TSDQ research.
With respect to state data bases, it is encouraging to know that their use in studying TSDQ issues continues to be valued. The work of individual states summarized in the Barro paper is indicative of such interest. The teacher common market project in the New England states and New York indicates interest in these issues not only at the state level, but also at the regional level. There is growing interest as well in the Southern and North Central states for regional data bases on teachers. In addition, the National Science Foundation clearly has an interest in issues of teacher supply and demand at the state level as evidenced by the State Science/Mathematics Indicators Project, which it sponsors at the Council of Chief State School Officers.
It is clear that a national analysis of teacher supply and demand without further disaggregation is inadequate. Our interest at NCES needs to go beyond this. Obviously there are subnational variations in this phenomenon, which we need to study. Even though our newest survey, the Schools and Staffing Survey (SASS), was explicitly designed to obtain state-level data, SASS is limited in fully exploring TSDQ issues at the state level because sample sizes will simply not support many of the needed analyses. Another advantage of state data bases is that many of them have a longitudinal component because states have been collecting teacher data for many years.
In conclusion, I will mention a few new NCES initiatives. Feedback from you on some of the actions we are currently planning to undertake and others that we are contemplating will be very useful.
With respect to SASS, we are now budgeted to readminister the Teacher Followup Survey (TFS) several times. Currently, SASS is designed as a
cross-sectional survey, complemented with the TFS, which after one year follows samples of teachers who (a) have left the profession, (b) transferred between schools, and (c) remained in the same school. Beginning with the 1992–93 SASS, we plan to follow these samples of teachers longitudinally as long as it proves fruitful to do so.
Another SASS innovation that we are pilot testing this year is the addition of a student component, the concept being to draw samples of students instructed by the individuals in the teacher sample. The feasibility and utility of this approach is now being explored. One of the advantages for including students in SASS is to understand better student course-taking patterns, one factor related to teacher demand.
NCES has also introduced (for fiscal 1989) collection of fiscal data in our state aggregate collection. This provides a much more powerful fiscal data base at the state level than we have had for at least 10 years. These new data should be useful for analyzing selected fiscal aspects of TSDQ.
We have also been working closely with the U.S. Bureau of the Census in collecting fiscal data at the school district level. Every five years, the Bureau's F-33 data collection entails a census of school districts, while in the off years these data are only state representative. In 1991, NCES provided additional funding to the Census Bureau so that the 1990 F-33 collection could be expanded to include all school districts. We will be able to link financial data from the district level with the many population characteristics generated by the census mapping project when information from the mapping project becomes available in late 1992 or early 1993, thereby enhancing the analytic utility of that data base. In the coming years. NCES will work closely with the Census Bureau in a number of ways to enhance the school district finance data.
NCES also has a new survey, the National Household Education Survey (NHES). The first administration of this survey was conducted in early 1992. Using a random digit dialing telephone survey method, we plan to survey a sample of 60,000 households. It is possible that the NHES will provide opportunities to explore some of the difficult issues in measuring the teacher reserve pool.
Choosing Teachers and Choosing to Teach
In February 1991, I was invited to address the opening session for the first presentation of a group of papers on the issue of teacher supply and demand. I was unable to attend that evening, but I have had a chance to review the work and to think about ways to consider the process of estimating the need for teachers and the supply of those willing to teach. I am grateful to Dorothy Gilford, who has continued to encourage me in this endeavor, as well as to Stephen Barro, Mary Kennedy, Richard Murnane, and others who have contributed to my thinking about this issue. I certainly share their concerns about the lack of adequate models to predict the elements of supply and demand.
I will leave to these expert authors the technical discussion of model development and of teacher quality and practice issues. Their papers are important and useful and will reward all who read them. I will concentrate on the issue of teacher quality and how it affects supply and demand.
Let me begin with an example of how standards of quality are thrown out the window when the demand exceeds the supply of quality recruits. In Baltimore a few years back, the district set out to identify top candidates for its teacher vacancies. It tested applicants, had them provide samples of letters they would write to parents, and asked them to respond to questions that teachers would ordinarily encounter in their work. A number of candidates were identified as unacceptable for hiring and received letters from the district to that effect. Then in September, the city hired the same candidates it had earlier identified as unacceptable. Why? Because after the supply of successful candidates had been hired, vacancies still remained. The district placed conditions on these ''Labor Day Specials,'' such as a
requirement that they go to school at night to improve their literacy and math skills. The point, however, is that in the face of an inadequate supply of teachers, the district immediately abandoned its interest in quality.
The Baltimore example is troubling because most of the rest of our nation's school systems would do, and have done, the same thing. We hear a lot about high standards for entering teaching, but even the modest standards our states now set are routinely set aside when supply lags. We hear a lot about teachers' knowing their subjects, but misassignment and out-of-license teaching are commonplace. In some countries, if there is an insufficient number of qualified teachers, students are sent home until a qualified teacher can be found. The custodial function of schools is compromised, not standards of teaching quality. Here it's just the opposite.
The future supply of teachers in our nation will come from the high school classrooms of America, where from 4 to 7 percent of students are achieving at the highest levels, as measured by various tests administered by the National Assessment of Educational Progress. Whether it is reading, mathematics, science, or other subjects just a scant few of our twelfth graders are demonstrating a level of knowledge and skills that is reasonable to expect of the majority of our prospective high school graduates—certainly the ones that go on to college, and that's half of them. That half is the source of our future supply of teachers. The highest achieving of these high school graduates are snapped up by our elite postsecondary institutions; few of them become teachers. Judging from the NAEP scores, the overwhelming majority of the students who go on to college are getting their high school education there. That's our future labor source, and that's the pool from which we'll be drawing our future teachers.
Contrast our situation with that of our competitors. In France, Germany, and other nations, the students who enter postsecondary education have passed very rigorous exams, far more demanding than anything you'll see on the NAEP tests. They send 20–35 percent of their high school students on to college, while we send 50 percent, but 100 percent of their college entrants have met standards that are at least as high as those attained by only 4–7 percent of our twelfth graders. You can safely say, then, that 100 percent of our competitors' future teachers come from the equivalent of our top group of students. Even if all our best students chose careers as teachers, the United States would still have a substantial quality gap to close.
Teaching in a nation as large as the United States is a mass profession, and it is hard to get uniformly high quality in a mass profession. And yet we continue to organize our schools on the basis of self-contained classrooms, which require large numbers of teachers and which mitigate against our avowed desire to have a qualified teacher in every one of them. Perhaps instead of consistently compromising on quality, we could experiment
with organizing work differently. Instead of one teacher per classroom—a teacher who may or may not be qualified—we could try differentiated teams of educators. The head of the team would be a fully qualified teacher who planned and taught the main lessons and who organized the work of other adults—interns paraprofessionals, etc.—who assisted groups of children in discussing and working through the lessons. The point of this illustration is not to endorse a model but to suggest that the demand question may have an organizational or structural solution that allows quality to be retained in the face of inadequate supply.
Another approach to the teacher quality issue might be pay for knowledge systems, similar to those found in industry. Currently, teachers advance on the salary schedule in part based on advanced degrees and credits. By and large, any postgraduate courses or programs are accepted. It's also the case that many teachers take courses to qualify themselves as school administrators, who get paid much more than teachers. There is no reason why employers and unions can't use the existing system to line up incentives in the right direction: Instead of rewarding any course taking, why not reward taking courses and completing programs that boost teacher quality? For example, most states do not require elementary school teachers to know much math and science, and, as far as I can tell, these teachers do not go on to take courses in these subjects (and how to teach them). Might this change if financial incentives were offered for them to do so?
As you can see. I view the issue of quality as both competency in subject matter and in how to teach subjects to youngsters. In this country, we've put methods over content and are generally satisfied if prospective teachers know content at about the level they will teach. But one characteristic of teachers in countries that are doing a better job of educating youngsters than we are is their high level of subject matter competency compared with that of U.S. teachers. Here, we value the sensitivity of the elementary school teacher above his or her knowledge and teaching skills in math. I fail to see how we can expect to have our students perform at world-class standards in math when we don't care whether or not our elementary school teachers know math. When we address the question of supply and demand, we must do so in the context of supplying those who clearly know the subjects we wish our youngsters to master and who can demonstrate that they know how to reach youngsters.
Now, on the brink of real shortages of teachers, is the best time for us to insist on a change in how we organize schools and in the requirements for subject mastery for prospective teachers. We are in difficult financial constraints. But rather than insist on business as usual, we should take our difficulties as an opportunity to move ahead with the development of new organizations for the schools of the next century.
You will see in the chapters of this book models for forecasting both
supply and demand for teachers. These models traditionally rely on fixed pupil/teacher ratios to determine demand. Supply estimates are quite often a function of counting those certified, and they ignore the questions of who remains in teaching, or ever teaches, and for how long. Supply and demand models also tend to ignore public policy issues.
Let me give you two examples of the kind of public policy issues that can wreak havoc with even the most sophisticated estimates of how many teachers we will need and of who and how many will be available to fill vacant positions. Take an issue like private school choice. Right now, most Americans are opposed to allowing public dollars to follow kids to private and parochial schools. But if public schools don't improve, pressure for privatization will probably increase and gain more converts. Consider the effect on teacher demand if large numbers of private schools grew up. Consider the effect on quality if their standards for teachers were even lower than those of the public schools, which they probably would be absent regulation.
Another issue that must be considered is the increasing cost of a college education and the decisions that are affected by those changes. Students who pay $40,000 and up to get an education must ask about the long-term consequences of borrowing money to get a teaching degree that allows them to go to work for low wages in a situation in which they have little autonomy and lousy working conditions. If we contrast that with other options available, we may find that, given the financial consequences alone, teaching will be a viable choice for fewer and fewer, regardless of the demand.
So interest in teaching is an economic issue, but it is also a social issue. We have seen evidence of this in the behavior of those who chose teaching as a career in the past. For example, social factors made teaching the occupation of choice for educated women and minorities who were barred from other professions. People chose teaching and did not choose the alternatives because of the conditions of their time, as well as their personal interests.
Political considerations can lead to changes in vocational interests and choices. During the Vietnam War, record numbers of people entered teacher education programs in order to get a draft deferment, and many have stayed to teach our children. Right now, we have an opportunity to take advantage of another political development to help schools get more teachers. Some look at the reduction in world hostilities and see a peace dividend in the form of retiring military personnel who may be interested in teaching. Tens of thousands of career military personnel may become available through programs jointly supported by the Department of Defense and the Department of Education. Similarly, many of the technical and professional staff
likely to be laid off in defense-related industries are promising sources of math and science teachers.
The point is that supply and demand estimates need to include or be analyzed in conjunction with the political, social environmental, and labor conditions that exist in the target and alternative occupations. Before college students invest in education as a career, they ask. "Will there be work for me when I graduate? Will that work be satisfying and earn me respect?" What kind of message are we sending to prospective teachers when pay is poor, prospects for advancement are lousy, and when the ratio of teachers to students is 1:18 in primary schools while the ratio of supervisors to teachers is 1:11?
Demand for services is one part of the equation; how many teachers we have to provide service is the other. This number will depend on such things as attrition, retirement, and the new people we attract to the profession. Attrition means, literally, the wearing down by friction, and I am certain that many teachers who have left the profession in recent years would find this definition apposite. Teachers have been worn down by increasing violence in schools, larger classes, overzealous supervision, and poor and dangerous school facilities. All these things take their toll and often lead some of our best teachers to cut short their teaching careers by a number of years. It would be difficult, if not impossible, to include these intangibles in equations of supply. However, we need to evaluate data in light of conditions teachers face and keep in mind the importance of fixing some of the conditions that drive teachers away.
We also need to pay attention to the new teachers who will need support systems if they are to survive. Studies tell us a number of troubling things about new teachers: that those with the highest credentials are the first to leave, that half leave within a short period of time, and that our present system of teacher induction fails 50 percent of those who use it. A 1990 report from the Department of Labor tracked education graduates from 1985–86. It reported that, in 1987, 7 percent of the graduates who had prepared to teach were out of the labor force, 2 percent were unemployed, 16 percent were working part time, and just 75 percent were working as elementary or secondary teachers. Imagine, though, if we were to create a support system for new teachers that allowed them to begin their professional lives the way young professionals do in other fields. They might have mentors, limited work assignments—rather than a full class load on the first day—and internships with a variety of school personnel. Under this system, we could save potentially excellent teachers who are now lost to the profession at the same time as we eliminated those who could not make the grade. Based on current projections of attrition at 5.6 percent of the teaching force, and our experience of hiring about half new teachers and half experienced teachers for vacancies, we could retain 22,000 more teach-
ers each year, increasing the supply and actually doing something about the quality issue as well.
Retirement is another important factor in teacher supply, and it will rise in importance as the teaching force ages. According to the Department of Labor, in 1988, 48 percent of teachers were over 40, and 18.5 percent over 50. By 1990, the number had grown by 83,000 to 19.9 percent over 50, with 1 in 10 being 55 or older. We will see an increase in the retirement rates as we near the end of the century, and we need to look on this as an opportunity to accelerate restructuring. We need to use the changeover to introduce new patterns of instruction and supervision and to focus resources where they can be of most use to children instead of simply allowing for incremental increases and decreases in our budgets.
Moreover, we must begin recruiting while young people are still trying to decide what career they will follow. Each increase of 1 percent in the retirement rate means we will need 30,000 more teachers. It means our current share of graduates who can qualify to teach, which may be 10 or 11 percent of all college graduates, must increase by 25 to 30 percent to accommodate the retirement increase. If an increase of that magnitude seems impossible under the present education system, then our consideration of supply models must include a discussion of the consequences of not changing.
I am pleased to have had an opportunity to offer my thoughts on these matters and especially honored to appear with people for whose work I have the greatest respect. Read on, and there will be a quiz at the end.