Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Monitoring the Supply Pool of Science and Mathematics Teachers In this chapter we look at the models used at the national and state levels for projecting the supply of teachers for the next year, the components of these models, and what we would like to know to monitor more effectively the supply pool and its components. The models examined in the panel's interim report (National Research Council, 1987c) were the projection model used by NCES and the teacher supply and demand models and projections developed in six states: Cal- ifornia, Colorado, Illinois, New York Florida, and South Carolina. They . . .. ~ ~ ~ J are most commonly projection models, which attempt to project teacher sun~lv and demand and reach conclusions about surpluses or shortfalls of teachers at a point in the future. Useful models should incorporate four major characteristics. The first is behavioral content, by which we mean models of relationships between variables in the environment and the behavior of actors in the education system. An example of a behavioral component in a model of teacher supply would be the estimated impact of salaries and working conditions on the decision of teachers to continue or to leave teaching. The models examined are limited by the lack of behavioral content. The second major characteristic of useful models is disaggregation by geographic area and subject field, and the third is quality measurement. Some of the models examined in our interim report incorporate useful refinements, such as the use of age-specific and field-specific attrition rates in projections of continuing teachers. But one key problem is a lack of useful geographic disaggregation. Moreover, only about half the state models examined disaggregated data by subject field). Nor do models deal in a satisfactory manner with the issue of quality. When models consider this dimension at all, the definition of a qualified teacher is equated with certification. 92
MONITORING THE SUPPLY POOL 93 Finally, useful supply models should include all the sources of supply. Among the state models and the National Center for Education Statistics model that were the central focus of the panel's interim report, none provides what we regard as a detailed analysis of the contribution of the various components of potential teacher supply. Most of the models ignore supply sources other than newly certified teachers or some equivalent. The NCES model until recently limited projections of new entrants to new graduates of teacher training programs. Other definitions that are used in state models include students enrolled in the state's education programs and the number of newly certified persons. NCES and some state models have more recently broadened the components of the teacher supply pool. Among these state models, the California PACE model (Cagampang et al., 1986) represents the most fully developed analysis on the supply side, with projections of the supply of new entrants from four sources: (1) new or recent graduates of California credentialing programs, (2) new credential holders from out of state, (3) teachers entering from the reserve pool of nonteaching credential holders, and (4) college graduates who pass the California Basic Educational Skills Test and obtain emergency credentials. Because of inadequate data sources and the lack of knowledge of the supply behavior of the various new entrant components, however, the PACE model relies largely on extrapolations of historical hiring patterns in the state, which are not the same as projections based on behavioral supply relationships Overall, it is the panel's view that current models of teacher supply and demand have very limited usefulness for defining education policy and consist of little more than plausible extrapolations of relationships that are largely based on cohort survival techniques on both the demand and the supply sides. None of the models has any serious behavioral content- i.e., on the relationship between changes in circumstances (e.g. salary, working conditions, pension benefits, economic recession) and changes in the numbers and kinds of people interested in obtaining teaching positions or in the numbers and kinds of teachers demanded by school systems. Since much of the research needed to incorporate behavioral content in supply models has not been done, the panel considers monitoring supply to be the best course of action. By monitoring supply we mean gathering data relevant to teacher supply periodically and monitoring trends in the data. In the short run, efforts are needed to improve the consistency, scope, and quantity of data available for monitoring teacher supply. Concurrent with monitoring, research should be conducted to support behavioral models. As research findings on the relation between the incentives discussed in the preceding chapter and supply become available and the relevant data bases are developed, resources can be devoted to behavioral modeling. This
94 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS chapter assesses the data that are now or could be collected to monitor the supply situation of science and mathematics teachers in this country. MONITORING POINTS ALONG THE SUPPLY PIPELINE The number of teachers employed in schools nationwide is augmented each year by new graduates from teacher training programs, newly certified teachers who enter teaching from other pathways (such as collaborative arrangements with industry), and previously certified teachers who are not teaching but have chosen to reenter the profession. The number of teachers employed is diminished by attrition due to retirement and other causes. Thus, monitoring supply requires keeping track of changes in the supply pool over its various stages. It requires data on certification, on incentives that motivate people to apply for or accept teaching positions, on new hires, and on attrition and retention rates. NCES recognized the importance of statistics to monitor teacher supply and demand in the United States and initiated the Schools and Staffing Survey (SASS), an integrated set of questionnaires that are designed to provide several of the types of information sought. SASS is described in greater detail in Appendix B. along with descriptions of other national data sets. The first SASS questionnaires (for school districts, schools, school administrations, and teachers) were fielded in school year 1987-88. A follow-up survey, of all teachers in the base year who left teaching and a subsample of teachers who remained in teaching (both those who remained in the same school and those who moved to another school), was conducted in the 1988-89 school year. Thus, not only are current teachers included in SASS; there is follow-up information from subsets of teachers who left and teachers who remained. If SASS produces the data sought, the survey will provide the most valuable data related to teacher supply and demand the nation has had. As with all new surveys, some skepticism is In order about the ability of SASS to meet all its goals. A useful way to envision components of supply that should be mon- itored is to identify stages along a pipeline, as outlined in Chapter 3. At the beginning stage are college students planning to teach. The pipeline progresses through degrees earned and certification, the decision to enter teaching, through retention and attrition rates. One could add high school students' aspirations to become teachers at the very beginning of such a pipeline. However, although a high school student's expression of interest in a future career generally indicates the de- gree of regard for that kind of career or calling, it Is probably not a reliable indicator of actual career choice. The Office of Technology Assessment's 1988 report, Elementary and Secondary Education for Science and Eng~neer- ing, usefully describes a pipeline model that includes precollege students'
MONITORING THE SUPPLY POOL 95 views of and preparations for science and engineering careers (O~ce of Technology Assessment, 1988:6-20~. It is used descriptively, however, and not for statistical modeling purposes. It states (p.6) that "students' inten- tions remain volatile until well past high school, with substantial numbers entering the pipeline (by choosing science and engineering majors) by their sophomore year of college." For purposes of monitoring data and gener- ating information on the supply of science and mathematics teachers that could possibly be used in models, we begin the pipeline at the college level. College Students Planning to Teach The proportions of students enrolled in postsecondary education who are majoring in education; in mathematics; in computer science; in physical, biological, and earth sciences; and in engineering are key components of the supply of science and mathematics teachers at this early stage of the pipeline. The number of education majors is in decline, and many who may wish to teach now pursue a subject major. Therefore, monitoring the number of education majors provides only a partial count of this component of supply. To gain a better understanding of the input to the pipeline, it would be desirable to make fuller use of the data that exist on freshman aspirations. Data from The American Freshman survey of the population of freshmen in higher education (described in Appendix B) include intended major and career aspirations and can be analyzed by sex and ethnicity as well. Data from the occasional follow-ups of those who have remained in college after two years and four years could be used to assess the value of freshman aspirations in judging changes in input to the pipeline. It would also be useful to have trend data on how many students who majored in the subjects noted above (subjects that are the most likely source of science and mathematics teachers) and who planned or did not plan to teach actually did or did not obtain certificates. The High School and Beyond longitudinal survey, which began in 1980, and the surveys of Recent College Graduates (both conducted by NCES and described in Appendix B) can provide these data for the science and mathematics majors in the samples of students who were surveyed. Research conducted by individuals using these data sets can address issues in focused and informative ways. For example, Maxwell (1986) used data from The American Freshman and followed a sample of 2,000 freshmen to their junior year, relating their intended majors (education, not education), intended careers (teaching, not teaching), and declared major as juniors to their high school grade point average and rank and college grade point average. His findings (that the group who had not intended to major in education but planned on a career in education and
96 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS later declared an education major had higher college grades than did other groups), although more related to quality, illustrate a wealth of information that could be used for research into supply questions, by subject major, at the postsecondary stages of a supply pipeline. In Chapter 6 the panel recommends measures to make these and other data more accessible to researchers. Certification Certification requirements vary from state to state, and the certification requirements in one state might not meet the requirements of another state. In some states a bachelor's degree that includes certain courses carries with it a teaching certificate. In other states, a year of teacher training beyond the bachelor's degree is required. In some states, as much as a master's degree (in teaching or in a subject discipline) is required for certification. In addition, 21 states now allow alternative certification routes (such as through a cooperative program with the military), and usually to staff particular subject areas with shortages (McKibbin, 1988~. The meaning of national data on the number of certified teachers is somewhat ambiguous because of the diversity of states' certification requirements. ~ quantify the pipeline leading to certification it is useful to look at the number of students enrolled in education programs. At present, the major source of such data is the American Association of Colleges for Teacher Education (AACTE), an association of approximately 1,200 member colleges and universities that have teacher education programs. Periodically the AACTE conducts surveys of small samples of member institutions to obtain data on the numbers of students enrolled in these programs. During the panel's May 1988 meeting with personnel officers of seven large school systems, they suggested that it would be useful to school districts planning recruitment of teachers to know the number of people in the pipeline by field. State education agencies should collect data on the number of students and graduates preparing for certification to teach, by field and by type of program (i.e., traditional or alternative) and make these data available to districts on a timely basis. These data are indicators at one juncture of the pipeline of potential additions to teacher supply within the next year or two. Changes in the actual number of graduates awarded certificates to teach science or mathematics should also be monitored. National data should be compiled from state certification board data on the number of new certificants by type (regular, alternative, emergency) and by subject annually. As noted above, although different states have different certifi- cation requirements and classifications, it should be possible to estimate
MONITORING THE SUPPLY POOL 97 comparable totals for categories such as science and mathematics (elemen- tary and secondary) and to present disaggregated data when available. Comparability across states would be more achievable if proposals to . . . ~ . ~ ~ . . ~ . , ~ ~ ~ standardize certification nationalb through board cert~catlon toescrloea in Chapter 5) are implemented. Since teachers may be certified in more than one category, these data will tend to overestimate the increase in the supply of newly qualified teachers. In fact, these data provide an upper bound of the change in newly qualified supply. Information is also needed on the degree of reciprocity in certification across states. This information can be used to indicate the extent to which shortages in one part of the country could be filled by additional teachers from another. The effect of reciprocity is a research issue that relates to the mobility of the reserve pool. The National Association of State Directors of Teacher Education and Certification publishes states' reciprocity provisions periodically in its manual on certification. Reciprocity, too, could cease to be a problem if current proposals for national board certification of teachers come to pass. Large proportions of new graduates of teacher certification programs do not go on to teach in the state in which they obtained certification. Why? Follow-up data on new certificants are desirable to ascertain the numbers and proportions of certificants who did not seek, were not offered, or did not accept teaching positions offered in their states. The follow- up questionnaire would probe for reasons why certificants did not teach, alternatives they pursued, and salaries. The survey of Recent College Graduates (RCG) is one possible source. It provides national data on graduates one year after receiving an education degree. The final survey in this series Is scheduled for 1991, but it will be replaced during the next two years by a national longitudinal survey of college graduates. States may also find it valuable to follow their new certificants to understand loss to the pipeline of teachers at this juncture. Data on the above aspects of certification would help to answer such questions as: How have increased state requirements for teacher certifi- cation affected enrollments in these programs or reentry after a gap in teaching? 1b what extent do states' restrictions and requirements placed ran t~.~.h~.r~ moving from another state discourage them from reentering ~ ~ ,^^_ · _^~ O teaching in the new state? Another occurrence that should be monitored as a possible indicator of shortage is the states' use of emergency or provisional teaching certificates, in science and mathematics in particular. According to the NEA survey Status of the American Public School Teacher 1985-86 (1987:20), only 8.4 percent of 1,291 respondents in all fields said they did hold such a certificate. (The data were not presented by subject area.)
98 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS Though the incidence seems minor, it would be desirable to have data from the states or local school districts, over time, on the number of teachers who hold temporary, provisional, or emergency certificates. SASS collects data from a sample of teachers on these types of certification in the teachers' primary, secondary, and best-qualified teaching fields. The design of the teacher sample will permit estimation of these rates of certification by type nationally and regionally and by type of district and school. Analyzed by subject, by region, or by type of area (e.g., rural, suburban, urban), these data may indicate exhaustion of the reserve pool or a shortage in a particular subject or in a particular geographic area. It would also be important to know how easy it is to convert a temporary certificate into a permanent one. Finally, it would be desirable to monitor the incidence of obtaining certification through various types of alternative certification programs. This could be done by adding a question on use of alternative certification routes to the questionnaire. New lIires Newly hired teachers come from many different sources, including new college graduates, former teachers, individuals who were certified but never taught, and teachers who change residence. School district personnel administrators indicated to the panel that typically a large percentage of their new hires were experienced teachers, not new certificants. Some administrators expressed a preference for experienced teachers. Descriptive statistics in many states indicate that a substantial fraction of new hires consists of teachers that fit into some category other than newly certified teachers. For example, less than 20 percent of new hires of mathematics and science teachers in New York State were new certificate holders; the corresponding figure in Illinois is 40 percent (National Research Council, 1987c:113,103~. An urban school district in a western state that took part in the panel's case study provides a more specific illustration of this phenomenon. There, only 4 of the 16 science and mathematics teachers hired in the past 5 years were new to teaching. And of those 4, only 1 came directly from college followed by a teacher training program; the other 3 had graduated between 3 and 10 years earlier and had recently gone back to school for their teaching certificate. Interviews with the teachers produced thumbnail sketches of previously experienced new hires: · A newly hired mathematics teacher had taught science for 12 years in a junior high school in a neighboring suburban district and grew to dislike it. After other jobs for four years, he chose to return to teaching but preferred mathematics to science to avoid the responsibility and liability of labs.
MONITORING THE SUPPLY POOL 99 · A newly hired chemistry teacher had taught for six years in a local parochial high school, following a career in the Navy where he did teaching and training full time. Without his retirement salary from the Navy, he would not be able to support his family on a teacher's salary. He was actively recruited by the district. · A science teacher, hired five years ago, had been laid off three times, twice in this district and once in another district, all due to seniority. With six years' teaching experience, he was offered a position in five schools in this district five years ago. · A mathematics teacher was hired on a part-time basis at her request. She had taught many years in another state, taught as a long- term substitute in this district, then as a regular teacher, and plans to return to full-time teaching next year after her children have adjusted to school. · A teacher was newly hired in mathematics but in fact had taught chemistry at the same school for seven years. She began as a mathematics teacher, which was her major, but after teaching 12 years in this district was slated to be laid off. So she switched to chemistry (her minor) and waited for a math position to open. · A mathematics teacher hired five years ago had taught at the junior high level in this district for eight years before moving to the high school level. These examples and information from other states underscore the need to obtain data on all of the components that make up the new teacher supply. Even new certificate holders may not be from the traditional new col- lege graduate channel: they may be older people who left their occupations to earn alternative certification. A total of 21 states offers such a channel, mainly to meet the needs of particular shortage areas. In most of these states, alternative certificants constitute a small percentage of total new hires (McKibbin, 1988~. But some states are notable exceptions (Carey et al., 1988:27-28~: The New Jersey Provisional Teacher Program trained 240 mathematics and science teachers between 1985 and 1988. Among the districts that participated in the California Teacher Trainee Program in 1984-85 and 1985-86, the program accounted for 61 (15 percent) of new mathematics teachers, 101 (31 percent) of new biological science teachers, and 24 (24 percent) of all new physical science teachers. 1b monitor the supply of science and mathematics teachers, it is necessary to be able to distinguish among the components of the corps of new hires. Ideally, to monitor supply at the district level, data are needed for the following categories by subject (chemistry, physics, biology, other sciences, calculus, other math, other subjects):
100 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS Certified teachers with no prior teaching experience: Certified before the last school year* Out-of-state certificate In-state certificate Certified during last school year Out-of-state certificate In-state certificate-whether through a regular or an alterna- tive program Certified experienced teachers: Teachers returning from temporary leave* Aught prior year Out-of-state In-state but out-of-district Did not teach prior year* Last taught out-of-state Last taught in-state but out-of-district Last taught in-district Noncertified college graduates with emergency credentials: Aught last year In-district Not in-district Did not teach last year* *These categories came from the reserve pool. Many of the districts in our case studies do not currently collect such data, and only a few (generally the larger districts) disaggregate the data by subject, as would be desired. SASS has been designed to collect such data. Several other types of data could shed light on the supply of new hires. These include information on incentives to teach, reasons for selecting one's current school or district, the number of applicants per opening, the number of job offers per hire, and the extent of district recruiting. SASS presently asks school administrators how difficult it was in general (not for specific fields) to find qualified applicants to fill teaching vacancies last year. If the respondent notes it was difficult in some fields, space is given to write in those particular subjects. Desired information regarding new hires includes the ability to monitor their incentives to teach. Comparative salary data are needed to indicate the competitiveness of beginning teachers' salaries relative to the starting salaries of alternative nonteaching positions. There is a question of just how this comparison should be made, but one simple measure would be starting salaries in industry for people with equivalent education, such as a bachelor of science degree in mathematics. The College Placement Council publishes these data annually (College Placement Council, 1988~.
MONITORING THE SUPPLY POOL 101 It would also be useful to know the reasons why science and mathe- matics teachers selected their current school or their current district and alternative offers they had. Such data are not currently collected by the SASS teacher questionnaire and were not part of the National Longitu- dinal Survey (NLS) teaching supplement. The data would help identify actions that schools or districts might take to attract well-qualified science and mathematics teachers. However, if there is a national shortage, such actions may only alter the geographic distribution of existing new teachers. Mend data on the ratio of applicants to vacancies by field could be useful, even though it is recognized that the number of applicants for a position is a function not only of supply but also of aggressive recruiting and of the characteristics of the district and its schools. Although an applicant may apply for more than one vacancy, a decline in this indicator over time would point to increasing shortages of applicants in a particular field. Questions on the number of applicants could be added to SASS. Mend data on the ratio of job offers per hire could alert a district to changes in the attractiveness of its positions, but small (large) ratios could also reflect a surplus (shortage) of teachers. Interviews with school district administrators revealed how shortages and surpluses in particular subjects were reflected in adjustments made in recruiting practices. A shortage in an area of need frequently for minority teachers at all levels-often would trigger an aggressive recruiting effort, including trips to other states. It was often mentioned that a few years ago officials traveled to cities experiencing teacher layoffs to recruit science and mathematics teachers; today, however, the officials may have few or no vacancies in these subjects, and no recruiting is needed. It would be helpful to have data from school districts, perhaps building on SASS, on the extent to which districts are shifting from screening applicants to recruiting actively by subject and by raciaVethnic group. When teachers are in surplus, districts recruit near home (if at all) and usually need only to screen and accept applicants. Thus, a count of the number of districts engaging in active recruiting is an indicator of shortage, and growth in these numbers over time suggests an increasing shortage. The Reserve Pool The reserve pool consists of people with teaching experience who did not teach last year, or individuals who were certified to teach at least a year ago but who have never taught. This reserve pool is a major source of new hires. The National Education Association estimates that more than half the new hires in the nation come from it (NEA, 1987f). In Connecticut, more than two-thirds of the new hires in 1984 came from the reserve pool. Returning experienced teachers constituted 55.5 percent of the new hires,
102 PRECOLLEGE SCIENCE AD ~THE~TICS TRACHEA and 12.0 percent were reserve pool members with no prior Connecticut public school teaching experience (Prowda and Grissmer, 1986:18~. Since the reserve pool plays such a major role as a source of new hires, it is important to know how large it is and where it is found, or at least to know whether it is nearly exhausted. Concern about depletion of the reserve pool is not a state issue, but can be a special urban, rural, or regional issue related to the region's particular labor market situation. In major parts of the state of Louisiana, for example, it is impossible to find certified teachers to fill vacancies. Consequently, the schools are forced to hire on a temporary certificate or to drop courses from the curriculum. In these areas, not only is the reserve pool exhausted, but there are also insufficient newly certified teachers who are willing to teach under the conditions offered. it, Little is known about who, among the various categories of people in the reserve pool, desires to enter or reenter the teaching profession. Thus, more important than estimating the size of the reserve pool is estimating the supply potential of the reserve pool, because some of the individuals in the reserve pool would not reenter teaching under any conditions. Different components of the pool can be expected to behave in very different ways. For example, teachers on maternity or health leave during a given year, or laid off and expecting to be called back, can plausibly be expected to return to the teaching pool in the next year at relatively high rates; newly certified teachers who did not obtain teaching jobs even though they have been in the market during the last few years, can be expected to remain in the teacher supply pool with relatively high probability; teachers whose credentials are older and who have been out of the teaching market for several years have a lower probability of being attracted back to teaching; while people with teaching certificates who have followed a completely different career path for many years have a much lower probability of being attracted to teaching. In some states, as we noted above, everyone with a bachelor's degree is potentially in the supply pool with some (arguably low) probability. Several approaches could be followed to estimate the supply potential of the reserve pool: new college graduates could be followed over time, new hires from the reserve pool could be tracked backward, data could be collected and accumulated on the number of last year's certificants who did/did not get teaching jobs, and state agencies could use state certification files to study the reserve pool in the state. Looking at these approaches in more detail, first consider what could be learned by following new graduates over time. The best source for this is data on the cohort of education majors in the Longitudinal Study of the High School Class of 1972. Heyns (1988) has used this data base to study entry and attrition of this cohort and to collect information on who left teaching and who wants to return (potential supply in the reserve pool). She
MONITORING THE SUPPLY POOL net 103 found that, by 1986, one-quarter of the education majors who completed teacher training programs had never taught. ONTO other longitudinal studies, High School and Beyond and NELS:88, will provide similar opportunities for studying the reserve pool in the future. Another source of one-year longitudinal data Is the Recent College Graduate (RCG) Surveys carried out by NCES periodically from 1976 through 1984. Using combined data from these surveys, the Office of Technology Assessment (OTA) analyzed the career paths of newly qualified teachers (1988:55~. It found that at about one year after graduation 20 percent of newly qualified teachers had not applied for a teaching position, and that the remaining 80 percent who had applied for teaching jobs could be separated into 51 percent teaching full time, 11 percent teaching part time, and 18 percent not teaching at all. In fact, one year after certification, 49 percent of the certificants were in the reserve pool. It is also possible to study new hires who come from the reserve pool, including both reentrants to teaching and individuals who have not taught but who were certified over a year ago. The SASS questionnaires for schools and teachers collect career information on both new entrants to teaching and experienced teachers regarding their recent career paths; these data will provide a rich resource for research on the reserve pool. It will be possible to study age of reentry, occupation just prior to reentry, the number of breaks in service, different patterns across different fields, and differences among different categories of members of the reserve pool. If the proportion of new entrants coming from different categories of the reserve pool (e.g., individuals certified in the past five years who are first-time teachers, returning teachers from the same district, returning teachers from the same state, returning teachers from out of state, etc.) are monitored over time, the trends will begin to shed light on the adequacy of the reserve pool to meet the demand. The SASS data are expected to provide an exciting opportunity for study of the reserve pool. At the state level, an important segment of the reserve pool consists of individuals certified by that state who are not currently teaching. By the use of data from the retirement files and the certification files, it is possible for some states to track certificates who still reside in the state and to characterize that segment of the reserve pool by age, certification field, and years of teaching experience. A survey of these individuals can determine interest in teaching or Incentives that would make teaching interesting to them. This would give some measure of the potential supply in this segment of the reserve pool. A study of this type was carried out in 1987 for the Commonwealth of Massachusetts by the Massachusetts Institute for Social and Economic Research (MISER). MISER's Report on the Status of Teacher Supply and Demand in Massachusetts (1987) used certification files to determine the
104 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS percent of those certified in Massachusetts in a given year who were not teaching the following yearns). Certificants were tracked for several years forward. It found that certificate holders were hired as long as 12 or 13 years after certification. Connecticut maintains an ongoing activity to monitor its reserve pool of teachers. Part of that effort focuses on the nearly 3,000 new certificants who did not take teaching positions within a year following certification. (Nearly 4,000 had received certificates; it was found that only 992 of them were employed in public schools.) A random sample of 36.4 percent of these individuals was surveyed in 1987 (Prowda and Beaudin, 1988~. Addresses were available for only 2,727 individuals certified in the state in 1985-86 but not teaching in 1986-87. Questions included current employment and salary, any prior teaching experience, undergraduate subject major and grade point average, teaching application experiences, and future employment plans. Certification files provided further information on subject specialties. Among the results: the median age of the nonteaching new certificates was 33; on average, they obtained their certificates five years after college graduation; slightly more than half had never taught full-time; the majority were employed, usually in education-related work (substitute teaching, enrolled as graduate students, tutoring, etch; 52 percent had applied for public school teaching positions for 1986-87; only about 30 percent of respondents intended to apply for the 1987-88 year; and proportionately fewer science and mathematics majors would apply than education, social studies, and humanities majors. Salary sensitivity was analyzed using the questionnaire data: "The probability that a non-teaching new certificant holding a job outside of education applied for a public school teaching position declined by 6.6 percent for each $10,000 he/she earned in salary" (P.91. The results of the Massachusetts and Connecticut studies are offered as examples of what can be learned about certificants who have never taught an important segment of the reserve pool. While national data are not available on certificants who have never taught, states can be encouraged to gather such data, and NCES can compile and disseminate states' information accordingly. No research areas concerning the reserve pool warrant study. One is to study how much change in incentives would be needed to attract reserve pool members to teaching. The Connecticut study on salary incentives could be repeated in other states. In addition, studying the number of returnees to school districts in which there have been large salary increases (e.g., Rochester, New York) could reveal some information on the size of the potential supply in the resolve pool. Another topic would be the effect of limited mobility on entry to teaching. It appears that people will move a short distance but not a long distance to accept a teaching position. A
MONITORING TlIE SUPPLY POOL 105 study of the recruitment areas for urban, suburban, small town, and rural districts would help define the geographic areas in which members of the reserve pool represent potential supply. In view of the major importance of the reserve pool to teacher supply and the concern that the other constituent of supply, newly certified teach- ers, is decreasing, the panel urges that all of these types of data collection and research activities be carried out. Retention and Attrition Rates New entrants to teaching are one major component of teacher supply. The other major component and the largest-is the corps of continuing teachers. Continuing teachers represent typically 90 percent of teacher supply in any year. Models tend to use single retention rates or attrition rates for projection purposes. More successful state models use attrition rates that are differentiated by age or years of experience and by subject field. We reaffirm our support, as stated in the panel's interim report, for the use of timely, disaggregated data to determine the proportions of teachers who can be expected to stay or leave. Improvements can be built into SASS to ensure useful data on teacher retention and attrition by subject. We again acknowledge that there are tricly problems in using information on retention to project continuing teachers. For example, teachers who leave one school may simply transfer to another, and for a national portrayal of supply this kind of mobility needs to be subtracted out. Models for a subject such as biology need to be sure not to count as continuing biology teachers those who were teaching another subject last year. In this vein, additional data are called for to monitor, over time, the important phases of the supply pipeline encompassing retention and attri- tion. 1b what extent do science and mathematics teachers leave teaching early in their careers? Do statistics indicate a large wave of retirements in the next five years? 15 years? It appears from an analysis by RAND of the teaching force in 1976-77, 1980-81, and 1983-84 that the proportion of teachers age 55 and over (9.5-10 percent) was quite stable. There was a clustering of secondary school teachers in mid-career, (age 35-44~; 36.2 percent of secondary school teachers in 1983-87 were in this age group, an increase from 22.6 percent in 1976-77. The older members of this group are expected to become eligible for retirement in about 15 years (i.e., 1998-99~. At that time the beginning of a wave of retirements may be anticipated (Haggstrom et al., 1988:8-9~. For science and mathematics teachers, pat- terns of age, years of experience, and expected retirements seem to be similar. The 1985-86 survey of science and mathematics education by Weiss (1987) found that the typical high school science or mathematics teacher
106 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS in 1985-1986 had 14 years of prior teaching experience, compared with 11 years for a typical teacher in 1977 some evidence of aging. The data do not support the prediction of an unusually large wave of retirees in the next decade. The class intervals in the Weiss study differ from those in the RAND study, although she found that only 14 percent of secondary school mathematics teachers and 16 percent of secondary school science teachers were over 50 years old (Weiss, 1987:64~. Data are thus needed from districts on the distribution of teachers by level, age, race/ethnici~, sex, and discipline. Attrition should be classified by retirement or other cause. Comparative salaries should be included to glean more about the competitiveness of teacher salaries relative to opportunity cost salaries. More specifically, information on the salary scale is needed, unless both average salary and years of experience are available. These data would enable the separation of attrition due to retirement from attrition due to incentives to leave teaching in favor of something else. The best prospect for obtaining these data is probably SASS, which included questions on attrition, by field, in the base year survey, but response to those questions was unacceptably small. The panel encourages efforts by NCES to modify the SASS matrix questions on attrition to simplify and improve response and to collect these data on a continuing basis. The school questionnaire should be able to separate attrition due to moving to another district from attrition due to leaving the teaching profession completely, which reduces the national supply of teachers. Overall changes in supply are affected by factors that make teaching more or less attractive compared with other occupations. The SASS follow- up survey of former teachers, conducted in spring 1989 and to be surveyed again in 1991 and 1993, should provide data on this aspect on a national scale. Another rich source of data that are available for analysis is the supplemental questionnaire sent to over 1,000 past and present teachers and those trained for teaching in 1986 as part of the fifth follow-up of the National Longitudinal Study of 1972 (NLS-72) (see Appendix B for a description of this subsample). In a preliminary analysis of the detailed career histories of these current and former teachers, Heyns (1988) found that nearly half (44.7 percent) of those who had taught for at least a year were no longer teaching by 1986. She found attrition rates to be particularly high in the first three or four years of teaching, and often it was male high school teachers who left. Most of those who left were single and took another job directly after leaving teaching; that is, the primary pattern of nonretirement attrition was not women leaving for homemaking. Another factor that can influence the retention rate for a school district or state is the portability of teachers' pensions. It would be useful to have comparative information from states on teacher retirement policies. Such information would be helpful for research relating retention rates,
MONITORING THE SUPPLY POOL 107 portability of teacher pensions and retirement policies, and particularly the sensitivity of attrition rates to the generosity of retirement provision. The erects and the relative strengths of incentives such as retirement systems and those such as salary and professional development opportu- nities on attrition or retention are worth pursuing through better data gathering. SASS does ask school districts about the minimum age, years of service, and penalty associated with their retirement plans. Nonethe- less, additional information is needed to improve the analysis of behavioral components at the retirement end of the supply pipeline. Teacher Mobility and Interstate Migration Little information is available on teacher mobility. Although teacher mobility in or out of a district affects supply at the district level, it affects supply at the state level only if the migration is interstate and affects national supply only if the teachers immigrate or emigrate. Effective monitoring of teacher mobility should start with separation data from the states, in the interest of avoiding double-counting. ~Ib be checked is whether the state can subtract "movers" from "leavers." Little is known of the effects of interstate migration on the supply of teachers for a given state, but large systems the panel interviewed in Texas, Nevada, and Washington are noteworthy for major recruiting efforts out of state, as are two of the large-city districts represented at the panel's May 1988 conference, the Los Angeles Unified School District and the school district of Dade Count, Florida. While some states may have information on the in-migration of teachers (see Able 4.1 for New York State), most states do not maintain such data; for those that do, the data may not be comparable with other states. The National Governors' Association (N GA) surveyed 15 states in 1987 for information on teacher mobility and teacher retirement system characteristics (NGA, 1988:9-10~. Six states were able to provide information on the number of new hires who had taught in another state. The NGA concluded: "Clearly, this is another example of the need for better educational statistics. Nonetheless, . . . the number of experienced teachers moving into a state is large enough to be of concern but manageable enough for a retirement system to attempt to accommodate mobile teachers." None of the states has data for out-migration to other states. One effort to overcome that deficiency is a project funded by the National Science Foundation through the Council of Chief State School Officers and the Regional Laboratory for Educational Improvement in the Northeast and Islands. Known as the Northeast Teacher Supply and Demand (NETSAD) study, it is a seven-state cooperative endeavor being undertaken by the Massachusetts Institute for Social and Economic Research. A primal
- u: c~ - o · - L. · - c, b4 4) cC c) o : o ~ o ~ · - a~ · - ~c ° ~ E g L. a, - v - o o CQ ·o 4 U] o n, U) 00 m 0 ~ 00 ~ ~ CD ut CO O . . . . . O O - .m bO L~ bO U. 00 108 L. ~Q CD ~ a O ~ C. CD C5) ~ ~ CO _. e~ ~ 14 0 . . · . . - oo o ~ ~ o o D CO CO ~ CO C9 C~ Ch e4 _~ tD C~ C~ CO O U. CD . . . . . CO C5) C~ ~ ~ oo oo ~ CO U. C~ ~ C~ CO o ._ Pc o o o ·~ ·- ~ n. C) - o CQ L4 0 O . - co ~ oq o ~ ~ ~ - o O cO -~ - - ~ oo ~ o o
~ ~ e~ ~ e~ O ct) {D a: ~ co . . . . . . . . . . . O tD C9 e~ ~ 0 co oo 0 ~ CD CO ~ O {D e~ oo ~e~ ~ CD a, ~ U~ CD e~ o~ ~0 e~ oo oo oo co co ~ a) ~ co c~ - - - ~ e~ ~ Cr] oo 0 ~ ~ O ~ O ~ {D O O 0 ~ ~ O co ~ O ~i CD O ~ ~ ~ o ~ ~o CD ~ oO ~ O oO C. e~ ~ 0 0 co e~ ~ u, 0 , ~ cO - - - o - ~ ~ ~ o . . . . . . a, ~ ~ ~ ~P 0 u~ e ~ O o, - c~ ~ ~ o ~ o o . . . . . . ~ o ~ ~ oo o ~ ~o cg ~ u, 0 ~ 0 co ~ e~ c~ u, (D 0 ~ e~ ~co {D ~ e~ C5) 0 o' u, ~ co - - - ~ ~ ~ co e~ ~ cn u) rY s~ o z ~ ~ E j ~ ~ ~ E ~ ~ o ~o o ~ - p L. o ._ IY U] L' o 3 . . CJ cO 109
110 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS concern of the project is to track the historical interstate migration of teachers within the region. It was borne of the interest of the Chief State School Officers in the seven states to establish a regional teaching certificate. The SASS teacher questionnaire includes information regarding the occupation of teachers just prior to their current positions, which should be a resource for studying mobility at both the state and district levels. In ad- dition, the SASS follow-up survey of teachers who leave the districts in the SASS sample, and further analysis of the NLS subsample of teachers and former teachers, should provide some useful data on the behavioral com- ponents of teacher mobility. Do they move for salary increases, for better working conditions, or because their family moves? A different perspective on the behavioral components of teacher mobility could be obtained from in-depth discussions with a group of district personnel officers. A series of in-depth conferences with school system officials is recommended in Chapter 6. To summarize the discussion of this section, the supply of precollege science and mathematics teachers can be envisioned as a pipeline marked by a number of stages or decision points. These decision points should be monitored for a clearer understanding of the incentives underlying the decisions made by individuals as they move through college and the world of work. Key decision points that call for further information are college students' selection of majors and career goals; earning certification to teach and to become certified in certain subject areas; the decision either to apply immediately to teach or to pursue a nonteaching activity; knowing about the composition of the group of new hires; understanding the supply potential of teachers in the reserve pool; and knowing more about the decision to remain in teaching or to leave. At each stage we have noted data that are needed to provide a clearer picture of the supply of science and mathematics teachers. A SPECIAL CASE: TTIE SUPPLY OF MINORITY TEACHERS In the interviews conducted with school district personnel administra- tors across the country, they frequently mentioned a shortage of minority teachers. If concentrated efforts were made to increase the number of minority science and mathematics teachers in particular, an indirect result certainly would be an increase in the overall supply pool of teachers. In this vein, then, we summarize the available data on the particular problem of the supply of minority science and mathematics teachers. Statistics collected by the National Education Association (NEA) indi- cate that the proportion of minority teachers fell significantly between 1971 and 1986. In 1971, 8.1 percent of the teaching force were black and 3.6
MONITORING THE SUPPLY POOL 111 percent were of other minority groups. By 1986, however, black represen- tation had dropped to 6.9 percent, and other minorities had declined to 3.4 percent (NEA, 1987e:14~. Comparisons with minority school enrollment proportions show how seriously under-represented minority teachers are. A policy statement issued in September 1987 by the American Association of Colleges for Teacher Education (AA(~-lE;) summarized data from the NCES on student enrollments (CES, 1987b) and from the NEA on teachers (NEA, 1987e) in a clear illustration of the problem (AAC--lL;, 1987a:3~: · Blacks represent 16.2 percent of the children in public schools, but only 6.9 percent of the teachers. · Hispanics represent 9.1 percent of the children in public schools, but only 1.9 percent of the teachers. · Whites represent 71.2 percent of the children in public schools, but 89.6 percent of the teachers. To trace the special problem of minority underrepresentation in pre- college teaching-and, when data are available, science and mathematics teaching in particular selected statistics are shown below to illustrate the monitoring of minority-teacher supply at eight stages of a supply pipeline. 1. Minority enrollment us higher education. The data on minority enrollments that follow were compiled by NCES and reported in its 1988 report Trends in Minonty Enrollment in Higher Education, Fall 1976-Fall 1986 (CES, 1988d). Black enrollments in higher education went from 1.03 million in 1976 to a high point of 1.11 million in 1980 and declined somewhat to 1.08 million as of 1986. Enrollment for other minority groups has increased steadily since 1976: Hispanics (from 400,000 to 625,000 in 1986) and Asians/Pacific Islanders (more than doubling from 200,000 in 1976 to 448,000 in 1986~. Black male enrollment has undergone the most significant rate of decline among minorities, having fallen about 7 percent between 1976 and 1986. In terms of proportions, minorities generally constituted approximately 17.9 percent of total enrollment in institutions of higher education as of lean increase from 15.4 percent of total enrollments in 1976. For blacks, however, the proportion had fallen from 9.4 percent in 1976 to 8.6 percent in 1986. 2. Interest in majoring in education. For college-bound black high school seniors who noted their intended college major on their SAT back- ground information forms, interest in majoring in education decreased between 1981 and 1984. Baratz (1986), citing the yearly profiles of college- bound seniors published by the College Board, reported that in 1981, 5 percent of black student respondents intended to major in education. But by 1984 only 3.4 percent of black students intended to do so (pp. 9-10~. And when the mean SAT scores of black high school seniors by intended major are compared, the highest mean SAT scores are found among students
112 PRECOLLEGE SCIENCE AND MATHEMATICS TEACHERS headed for the fields of engineering, physical science or mathematics, and biological sciences. The mean SAT score for students who were planning to major in education was the lowest among the eight fields of study (Baratz, 1986:8~. 3. Bachelor's degrees earned. It appears that the minority teacher supply pool continues to decline with each stage along the pipeline. Most of the declines occur among blacks; other minority groups seem to continue in the supply pipeline at greater rates than blacks, and at greater rates than a decade ago. As reported by NCES in Education Indicators- 1988 (NCES, 1988f:104~: Blacks earned fewer degrees in 1985 than in 1977 at all degree levels except the first-professional (e.g., M.D., J.D.~. Ibe declines are particularly significant when compared with increases in the young adult black population during the same period: it rose 7 percent among 18- to 24-year-olds and 40 percent among 25- to 3lyear-olds. Men accounted for nearly two-thirds of the drop in degrees [among blacks] .... Hispanics, Asians and American Indians/Alaskan Natives earned more degrees in 1985 than in 1977 at all levels. The increase among Hispanics was in line with their population growth. The number of bachelor's degrees in education (though this statistic has limited meaning given the movement toward requiring subject majors) fell significantly overall, from 143,462 in 1976-77 to 87,788 in 1984-85 a 39 percent decline. But for blacks it declined particularly steeply. In 197~77, 12,922 blacks earned bachelor's degrees in education; in 1985, only 5,456 did a 58 percent decline (NCES, 1988f:290,292~. 4. Pursuing a master's degree in education. In an NCES special report, Hill (1983:18) reported a large decrease between 1976 and 1981 in the number of education degrees earned at the master's and doctor's degree level as well as the bachelor's degree level. Fewer students as a whole obtained an advanced degree in education, but for blacks the number declined more than for graduates in general. Overall, 128,417 graduate students received master's degrees in education in 1976 (NCES, 1986:130), by far the most popular advanced degree pursued (next was the master's in business, with 42,512 recipients). By 1985, only 76,137 master's degrees in education were earned a dramatic 41 percent decline (NCES, 1988b:211~. For blacks, however, there was a 53 percent drop in~the number receiving master's degrees in education: from 12,434 in 1975-76 (Hill, 1983:27) to 5,812 in 1984-85 (NCES, 1988b:223). 5. Pursuing teaching versus other endeavors. While the number of blacks earning bachelor's and master's degrees in education fell substan- tially between the mid-1970s and the mid-1980s, the total number of
MONITORING THE SUPPLY POOL 113 first-professional degrees and M.B.~s awarded to blacks rose during this period (Hill, 1983:27,29; NCES,198~3b:223,228~. In a Phi Delta Kappan article documenting the shortage of black teachers, Patricia ~ Graham cited a vanes of data sources indicating decline in the supply of minority teachers. Graham speculated about the alternative routes black students In higher education could be taking (Graham, 1987:603~: Experts differ about the causes for the decline in the number of black teachers and for the decline in the number of black students seeking to major in education. The first explanation, favored by those who believe that America is slowly but inevitably progressing toward racial justice, stresses the broader range of career choices available to educated blacks today. Previously, teaching was one of the few jobs available to college-educated black men and women. Because we now see blacks in other professions, we conclude that these successful middle-class blacks would have been teachers in previous generations. To some extent, this conclusion is probably correct, but not enough blacks have moved into other professions as yet for us to be certain that they represent displaced teachers .... Young blacks who are choosing alternatives to careers in education are not shifting in significant numbers to other professional fields. The data we have available do not probe the alternative careers that minorities are pursuing, and we believe this will be an important research topic to study. The Survey of Recent College Graduates asks for the respondent's race. This survey may hint at decisions made immediately after college graduation, but it does not probe as deeply into alternative career or schooling decisions as would be required to analyze fully this special case of minority science and mathematics teacher supply. Data from the follow-up surveys of the NLS-72 sample could be analyzed by race/ethnicity to identity the career patterns minority students are selecting and the career patterns they pursue. 6. New hoes. The number of new hires into the nation's public school systems is not known nationally by raciaVethnic group, but it should be available from SASS since the teacher questionnaire asks for the re- spondent's race and subject taught. SASS is also expected to provide data on what the person had done before entering teaching (though it does not ask for the salary of the job held prior to entering teaching). 7. The current teachingforce. Among the general public-school teach- ing force, the percentage of teachers who are black decreased from 8.1 percent in 1971 to 6.9 percent in 1986 (NCES, 1988b:70~. The proportions Of secondary science and mathematics teachers in 1985-86 who are black are lower still. Among grade 10-12 mathematics teachers, only 3 percent are black; among grade 1~12 science teachers, 5 percent are black (Weiss,
114 PRECOLLEGE SCIENCE AND MATHEAL4TICS TEACHERS 1987:63~. Most of the black and Hispanic teachers of science and mathe- matics are in the elementary grades. The Office of Technology Assessment (1988:58) posits: "For now, the proportion of minorities in the teach- ing force is increasing slightly, but several commentators warn of future shortages of minority teachers, particularly in mathematics and science." 8. Retention and attrition. Of 308 minority teachers surveyed as part of Metropolitan Life's 1988 Survey of the American Teacher, 40 percent said they were likely to leave teaching within five years, as opposed to 25 percent of the 891 nonminority respondents (Metropolitan Life, 1988:22~. This may reflect the fact that 29 percent of the minority teachers surveyed for the Metropolitan Life project worked in inner-city schools, as opposed to only 9 percent of the nonminority teachers. The Metropolitan Life report summarized its findings on this aspect of teaching as follows (p. 5~: Almost three out of four of the dissatisfied minority teachers say they are likely to leave, compared to about half of the dissatisfied non-minori~ teachers. Even among minorities who are very satisfied with their careers as teachers, more than one out of five say that they are likely to leave. · Less experienced minority teachers are the most likely to say that they will leave. Fully 55 percent of minority teachers with less than five years of teaching experience say that they are likely to leave the profession. Little information is currently available on retention and attrition rates of minority teachers. A study by Kemple (1989) of the career paths of 2,535 black teachers in North Carolina finds that, on one hand, over two- thirds of these teachers, who began their teaching careers between 1974 and 1982, stayed in teaching through 1985-86, and over one-third who had quit returned. On the other hand, the likelihood of teachers in this group leaving has been increasing since the mid-1970s, suggesting that as blacks gain access to other professions they may shorten their teaching careers. Kemple stressed that "even minor increases in attrition will have large influences on the overall representation of Black men and women in the teaching force" (p. 2~. Data from the 1989 SASS follow-up questionnaires of teachers who left and teachers who remain, which are expected to be available in 1990, and analysis of the 1986 survey of teachers and former teachers who took part in NLS-72, offer the best possibilities at this time of national data capturing much-needed information on alternative jobs taken and opportunity costs associated with teaching science and mathematics. These questionnaires ask for the respondent's race, subjects taught, length of experience, comparative salaries, and decisions made regarding continuing or leaving teaching, the pursuit of more schooling, or taking other positions.
MONITORING THE SUPPLY POOL 115 Research is needed on the employment patterns of minorities and especially the difference in response to opportunity by minority women. Would increasing salaries attract them to teaching? Or is the cost of teacher training the problem? From 1958 to 1965 state teacher colleges provided blacks with easy access to education. They could live at home while going to college and could get a degree with the opportunity for income. A researchable question is whether the cost of teacher training relative to other occupational training has increased. In conducting such research, analysis of existing data by raciaVethnic characteristics would be a productive first step. As a companion to the description in Appendix B of national data sets relevant to teacher supply and demand, NCES has published a useful compilation identifying six minority student issues and relating them to 32 NCES surveys containing racial/ethnic data (NCES, 1989b). In conclusion, the troubling evidence thus far on minority science and mathematics teachers suggests a disproportionately acute shortage of blacks. The most valuable activity we can suggest from analysis of data from SASS and other data sets and through the conduct of further research would be to probe into alternative decisions black college students and graduates make, alternative positions they pursue, and the opportunity costs they perceive that draw them away from teaching science or mathematics. SUMMARY We have called attention to important gaps in the content of current state and national models. 1b understand fully the forces, influences, and incentives that affect the supply of science and mathematics teachers requires data and behavioral content that the current models do not capture. A sequential approach toward the goal of improved national models is thus recommended. For the short term, efforts can be made to monitor the state of supply-by further analyzing existing data such as those from the NLS-72, by building on the promising work of SASS, and by compiling and disseminating states' data, for a clearer portrayal of the supply situation in this country. This chapter described existing and proposed data that can serve as components of an effective monitoring system.
