Cover Image

PAPERBACK
$58.25



View/Hide Left Panel
Click for next page ( 168


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 167
~ Bow William P. Butz, Gabrielle A. Bloom, Mihal E. Gross, Terrence K. Kelly, Aaron Kofner, Heiga E. Rippen Science and Technology Policy Institute RAND This paper has the following objectives: To clarify the concepts of "shortage" and "low production" in the context of scientists and engineers To suggest answers to the questions in the paper's title To point toward strategies for increasing the science and engineer- ing (SHE) workforce WHAT WOULD A "SHORTAGE" OF SCIENTISTS AND ENGINEERS LOOK LIKE? Over the last half-century, numerous alarms have sounded about looming shortages of scientists and engineers in the United States. What is meant by "shortage" has not always been clear. Further, the population under discussion, the scientists and engineers themselves, has not always shared the perspective of those sounding the alarm. Regardless, the im- plications of a shortage of skills critical to U.S. growth, competitiveness, and security are significant. So are the implications of the continuing low entry of female and minority students into many SHE fields. These impli- cations justify closer examination of the nature and sources of the over- or underproduction of scientists and engineers. Improved understanding of Note: The views expressed here are those of the authors and not necessarily those of the Science and Technology Policy Institute of RAND.

OCR for page 167
PAN-~CANIZAHONAL SUMMIT the definition and nature of the problem can point toward relevant data and useful questions. As a starting point, consider the different circumstances in which the production of any good or service, new S&E Ph.D.'s being one, might be called "low": 1. If production is lower than in the recent past (steel is a recent ex- ample) 2. If competitors' share of total production is growing (electronic com- ponent manufacturing, shoe manufacturing, and oil production are in- creasingly foreign) 3. If production is lower than the people doing the producing would like (automobiles, best-selling novels, blockbuster movies) 4. If less is produced than the nation is deemed to need (well-trained K-12 teachers, community volunteers, clean urban air) 5. If production is not meeting market demand, as indicated by a ris- ing price (nurses, Washington-area housing). Each of these concepts of "shortage" has a place. The pain of steel workers and their communities is real when production falls and plants close (concept #1~. The nation's concern about reliance on Mideast oil is justified (concept #2~. And so forth. However, one of these five concepts of "shortage" is fundamentally different from the others in a manner cru- cial for the question at hand. Only the fifth concept integrally embodies a corrective mechanism that solves the problem, that induces increased pro- duction of its own accord. To see this, consider the S&E workforce. If production of scientists and engineers is insufficient to meet market demand that is, if each new crop of American scientists and engineers is too small to fill the growing number of jobs offered by academic, industrial, and government employ- ers then salary offers will tend to increase and unemployment or under- employment of the S&E workforce will tend to diminish. As young people observe this tightening labor market and consider lifetime employment prospects along with the many other factors influencing their career choice, some of them will opt for S&E, rather than for clinical medicine, law, business, or another profession. As these people complete their edu- cation and join the workforce, total production of scientists and engineers will accelerate. The shortage will diminish. To the extent that production is "low" in any sense other than this fifth sense, production will tend to stay low. For example, the fact that competing countries are manufacturing more electronic components while America produces less (concept #2) may constitute a "shortage" of American-produced components. But there is nothing about this kind of

OCR for page 167
shortage that will induce American companies to reverse the move off- shore. Indeed, in whichever other respect there is a "shortage," policy ac- tions to relieve it will be effective to the extent that they operate to in- crease demand for the good or service (concept #5~. Policy can also induce increased production by lowering the costs of production, regardless of the manner of shortage that exists. IS THERE A SHORTAGE OF SCIENTISTS AND ENGINEERS? Diverse data from the National Science Foundation, the RAND RaDiUS database, the U.S. Census Bureau, the U.S. Bureau of Labor Sta- tistics, the National Research Council, and scientific associations can characterize the production of S&E Ph.D.'s, indicate the respects in which such production may be low, and point to causes of observed patterns. Accordingly, we briefly focus such data on the five different concepts of "shortage," indicating the particular respects in which the production of S&E Ph.D.'s indeed appears to be low. This overview points to the fifth concept, unsatisfied demand, as the key both to un- derstanding and to correcting whatever shortages are thought to exist according to the other concepts. Unfortunately, the uneven detail, varying definitions, and inconsis- tent time periods in the available data make possible only the teasing out of "stylized facts" hypotheses awaiting empirical testing. That data more recent than 1999 or 2000 are generally not yet published is especially un- fortunate, as the S&E workforce situation has arguably changed signifi- cantly since then. Hence, we conclude this analysis not with positions or solutions, but more modestly, with four possible strategies for increasing the production of S&E Ph.D.'s in whatever fields might be deemed low, by whatever criteria. To begin, consider whether the United States is experiencing a short- age of S&E Ph.D.'s in either of the first two senses decreased production or gains by competitors. Figure 1 shows that the number of American Ph.D.'s awarded in each major area of science and engineering has been increasing, beginning in the 1980s. These gains were interrupted in the late l990s, an interruption that has apparently continued in some fields, although confirming data are not yet published. Hence, at least until very recently, American Ph.D. production has not been declining in the broad S&E fields. There has been little or no shortage of the first type. What about the second concept of shortage competitors gaining ground? Figure 2 shows that S&E doctorate production turned up in many other countries during the 1980s as in the U.S., but the numerical increase has been larger here than in any of these "competitor" countries.

OCR for page 167
170 9,OOO 8,OOO 7,000 6,OOO to 2 5,000 us 4,000 o ~. 3,000 E Z 2,OOO 1,000 o PAN-~CANIZAHONAr SUMMIT _ I_ t- :::::: I ~ Social and Behavioral Sciences -X- Mathematics l ~~ ~ Physical end Geosciences ~ Engineering _ _ Biological and Agricultural Sciences - ~ Computer Sciences OF ~--X--~- ~__~__3d,~ ~~ ~-~~ y ~ ~ ~ ~ ~ ~ ~ ~X in-- - ---- ~ ~ A--- 1 1 1970 1975 1980 1985 1990 1995 2000 2005 FIGURE 1 S&E Ph.D. Degrees Awarded by Broad Field, 1975-1999. Source: Science and Engineering Doctorate Awards (20024. The American Bar Association (2002~. From two other angles, however, the situation vis-a-vis our "com- petitors" does not appear so sanguine. Figure 3 reports the ratio of S&E first degree holders to the total population of 24-year-olds in the so-called G7 countries in 1975 and 1999. Think of the height of each column as rep- resenting the probability that a representative young person will com- plete an S&E degree. That probability for American youth grew from .04 in 1975 to .06 in 1999, a notable increase corresponding to the numerical growth evident in the first two figures. In 1975, this probability in America was exceeded only in Japan, among countries shown. After 1975, how- ever, the picture is radically different. Each of the other countries has ex- perienced a much larger increase, measured in either absolute or percent- 30,000 25,000 20,000 ~ 5,000 - ~ 0,000 5,000 o ~ X ~ ~ _ ~ | ~ France -X- United States . , Japan | _ ~ Germany ~ China ~ S. Korea ~ United Kingdom ~ India Taiwan a* -eke >~ an__ ){,__-X~ { __~~ ~ _ = ~ ~ ~~ I,. ~ 'by 1 1 1 1 1 1 1 1 1 1 1 1 ~,_, - . At, - - _/ :*~'' - _~ _' 3~. - _ _ S. ~ By: ~ ' '~'~~~ -I' 1 ~ ' 1 ' '' '; 1 ' ''1 '> ' ' ''1'' ' '75 '76 '77 '78 '79 '80 '81 '82 '80 '80 '85 '86 '87 '88 '89 '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 FIGURE 2 S&E doctorates awarded in nine countries, varying years 1975-1999. Source: Science and Engineering Indicators (2002J.

OCR for page 167
- iL u] ~7 . to ~ 8 2 O >a ~ . es to %P E N ~ ' iL o fl) ILL . ~ 10 _ 6 4 2 O 17:! 1O 1 1~1 1975 1999 1 . l l 1 _ 1 1 1 1 1 United States United Kingdom France Japan Canada Germany Italy FIGURE 3 Ratio of natural science and engineering first university degrees awarded to 24-year-old population, by G7 country, 1975 and 1990. Source: Science and Engineering Indicators (20024. age terms.1 Although their young-adult populations are growing less rap- idly than ours (not shown), the proportion of their young people opting for university degrees in science and engineering is rising faster. Figure 4 examines from another angle the question of whether our competitors are gaining. Here, doctorate recipients from American insti- tutions are divided into U.S. citizens and non-citizens.2 The latter's share has grown rapidly indeed, from 23 percent of the total in 1980 to 42 per- cent in 1994. Even with the subsequent decline in noncitizen degree awards,3 this longer-term rise, combined with the increasing propensity of students abroad to enter S&E fields (Figure 3), buttresses the case that the American S&E workforce is low in the sense that our competitors are gaining (concept #2 above).4 Consideration of the third and fourth concepts of "shortage" is best deferred until we have taken up the fifth and last concept: Is growth of the S&E workforce insufficient to satisfy market demand? If such growth is insufficient; that is, if the numbers of American scientists and engineers are too small to fill the new jobs offered by academic, industrial, and gov- ernment employers, then employers will be bidding to fill their empty positions. Job openings, lab facilities, salaries, advancement opportuni- 1Other countries can be seen in the source table to have experienced even larger increases, notably Mexico and Spain. 2The underlying data for this figure include M.D.'s. Permanent residents are counted with noncitizens. 3Since September 2001 the number of foreign students enrolled in graduate S&E programs in the U.S. has apparently decreased even more markedly. 4Of course, many foreign recipients of U.S. degrees choose to remain and work here.

OCR for page 167
172 1 8,OOO 1 6,OOO 1 4,000 12,000 2 ~ 1 O,OOO ED ~ 8,OOO o E 6,OOO at 4,000 2,OOO o PAN-~CANIZAHONAr SUMMIT - / ~ | ~ U.S. Citizens Non-Citizens | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 FIGURE 4 S&E and Health Doctorates Earned by U.S. Citizens and Noncitizens, 1980-2000. Source: Science and Engineering Indicators (2002J and Science and Engi- neering Doctorate Awards 2000. ties, and other components of career satisfaction will be on the rise, while unemployment and underemployment will be falling. Alternatively, if the rewards to other careers perhaps clinical medi- cine, law or business are higher and are growing relatively to S&E, and if it is instead the costs of training for a job that are growing for S&E, then there is no shortage of scientists and engineers in this important fifth sense. Indeed, in this latter case, the "shortages" that others discern may well look more like discouraging surpluses to young people con- sidering career choice. In the sense that matters for spurring production, they indeed are. Is there in fact unsatisfied demand for scientists and engineers in the American job market? Available data are sketchy but they are consistent. We consider two indicators of S&E career opportunities: earnings and unemployment. Where data allow, we compare these opportunities and costs to those facing budding holders of professional degrees M.D., D.D., D.V.M., T.D., and M.B.A. These comparisons are instructive to the extent that bright ambitious young people consider other challenging alterna- tives while deciding whether to become scientists or engineers. Figure 5 compares a measure of annualized earnings for Ph.D.'s (all Ph.D.'s are included in this measure, not just S&E)6 with earnings of pro- 5Prevalence of postdoctoral appointments, particularly successive appointments, might be considered an indicator of underemployment. 6These highly aggregated data cannot reveal salary trends for just the S&E workforce, much less for particular disciplines and subdisciplines that may have experienced unusual salary growth or decline. For comparison purposes, about 60 percent of Ph.D. degree hold- ers were in S&E fields in the period covered by these data.

OCR for page 167
20O,OOO 1 8O,OOO 1 6O,OOO 1 4O,OOO us a, a, 1 2O,OOO IL lo 1 0O,OOO ~0 ~ SO,OOO o 6O,OOO 4O,OOO 2O,OOO o ]7L] - - | ~ Professional ~ Doctoral | 25 - 29 30 - 34 35 - 39 40 - 44 45 - 49 50 - 54 55 - 59 60 - 64 Age Groups FIGURE 5 Synthetic estimates of work life earnings for advanced degree holders by age, 1997-1999 period. Source: Census The Big Payoff. fessional degree holders (those listed just above).7 Professional degree holders earn more at nearly every age and considerably more over an entire career, as measured by the summed difference between the lines. ~ ~ . . . . .- IS IS no surprise. Our purposes here would be better served by this same earnings mea- sure calculated separately for the S&E workforce and repeated for a de- cade or so earlier. This comparison would reveal whether the professional degree premium is falling; that is, whether the relative attractiveness of an S&E career is rising, indicating a shortage in that crucial fifth sense. Alas, this measure is not yet available separately for S&E or for earlier periods. Still, the data at hand give no indication of the kind of earnings premiums for scientists and engineers that would signal the existence of a shortage. Unemployment rates are another indicator of market conditions. Rates that are falling or lower than in alternative occupations also suggest shortages in the fifth sense unsatisfied demand. Unemployment rates are available and plotted in Figure 6 for chemists, recent mathematics 7Called by the Census Bureau "synthetic estimate of work life earnings," this measure calculates for the 1997-99 period the annual earnings of persons in each indicated age range. A young person today might interpret the lines connecting these age points as the expected career profile of annual earnings on into her future. That interpretation requires several strong assumptions. An alternative measure of the career earnings profile would report an- nual earnings of the same group of people as they age over the years. As those data must necessarily refer entirely to the past, even to the deep past when the group of people was young, they also are a flawed proxy for looking at the future. However, lacking real data about the future, people and organizations use information about the past and present to make decisions, including career decisions.

OCR for page 167
174 12 10 E 6 PAN-~CANIZAHONAr SUMMIT Recent Math Doctorates Overall U.S. Recent Biomedical Doctorates ~ Chemists / \ 4 2 o 1 970 1 975 1 980 1 985 1 990 1 995 2000 2005 FIGURE 6 Unemployment rates of the United States and selected S&E fields. Sources: ASM-IMS IAAA Annual Survey 2001. ACS Annual Salary Survey. Chemi- cal ~ Engineering News 77: 28-39. NRC Trends in the Early Careers of Life Scientists. Ph.D.'s, and recent biomedical Ph.D.'s and M.D.'s.8 Although not fully comparable in population or time period, these three rates, when com- pared to the overall U.S. unemployment rate, suggest a general increase or leveling in the l990s, while the general unemployment rate was falling substantially. Rising unemployment in one sector, while the overall economy is doing well, is a strong indicator of developing surpluses of workers, not shortages. Hence, neither earnings patterns nor unemployment patterns indi- cate an S&E shortage in the data we are able to find. Altogether, these data in Figures 5 and 6 do not portray the kind of vigorous employment and earnings prospects that can be expected to draw increasing numbers of bright and informed young people into science. We return now to the third concept of shortage: Is production lower than the people doing the "producing" in this case the young people making career choices would like? More young people today may argu- ably enjoy doing science or engineering than plan actually to prepare for such careers. Instead they may choose a professional degree but only re- luctantly. In a market economy, even one characterized by rigidities, regu- lations, and unequal opportunity, qualified people tend toward career paths whose rewards and satisfactions are becoming more attractive and/ or whose preparatory costs are becoming less onerous. The American Mathematical Society and American Chemical Association publish more extensive data (including unemployment rates) on their members than are available for most other S&E communities.

OCR for page 167
8.5 8 7.5 a) 7 o 6.5 6 5.5 5 .~ _ _ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 FIGURE 7 Average registered time to Ph.D. in the biomedical life sciences (postbaccalaureate study). Note: Includes all graduate education. Source: Trends in the Early Careers of Life Scientists (1998~. We have seen that broad fields of science and engineering do not ap- pear particularly attractive from the earnings and unemployment perspec- tive. What about the cost side? Figure 7 points to a sobering part of the answer. Average time from bachelor's degree to Ph.D. in the life sciences has increased by two full years since 1970. Professional association staff in several other sciences informally confirm similar increases. If, as is likely, the variance in time-to-degree has increased along with the mean, then prospective life scientists face not only more years out of the labor market, but also more uncertainty about the number of years. Complaints about perceived subjectivity and arbitrariness of the postgraduate process its length and the prospects of eventual completion are also not infrequent. All this might not matter so much if the brightest young people lacked alternative training and career paths. But consider the paths to the M.D., D.D., D.V.M., T.D., and M.B.A. The number of years to degree has stayed absolutely constant in these programs for decades,9 and the prospects for successful completion, once begun, remain high. Have the amount and complexity of material to be mastered expanded so much more in biology or mathematics than in medicine, the law, or finance? This would seem hard to argue. Then why does it take longer and longer to be ready to begin one's career in most of the sciences, but not in the professions? Finally, what about the fourth concept of shortage, unmet national needs? Will particular subfields of science or engineering soon become critical, perhaps for national security, for health care, for feeding the world, or for national competitiveness? Perhaps some fields are already 9Flexible training alternatives that can extend time to degree have arisen in each of these fields, but these are optional, serving to increase the attractiveness.

OCR for page 167
i76 PAN-~CANIZAHONAL SUMMIT critical but somehow without the corresponding inducements that attract qualified young people. Where would these inducements come from? GENERAL STRATEGIES FOR INCREASING THE SCIENCE AND ENGINEERING WORKFORCE We have seen that the production of American scientists and engi- neers is not low in the sense that it has fallen over some years from previ- ous heights, nor in the sense that employers are driving S&E earnings up and unemployment rates down in a scramble to hire more. However, in another sense of shortage that of competitive foreign gains American production does appear low. Whether from unmet national needs, foreign competition, or any other source, a perceived shortage of U.S. S&E talent must be expressed in terms that motivate young people, or the shortage will persist. If such perceived shortages do emerge, the story loosely told by these data points toward four general strategies to relieve them. Two of these strategies involve government actions to increase the returns and rewards to be expected from a career in science. The other two strategies somewhat less ame- nable to direct government policy would reduce the costs of preparing for such a career. A fifth strategy points to data improvements. 1. Steadily and predictably increase federal research obligations for the S&E fields of concern. This strategy, though not easy, is straightforward. There is nothing so directly under control of the federal government as its bud- get, and probably little that has so direct an effect on the attractiveness of an S&E career. Federal grants, contracts, and other S&E expenditures are a major determinant of fellowship support, job opportunities, lab facili- ties, and salary growth. Figure 8 shows that federal obligations for total research, in constant dollars, have increased more than fivefold since 1970 in some fields, but hardly at all in others. The substantial growth in fed- eral support for the biological sciences seen in Figure 8 is likely a major reason for the corresponding growth in biological science Ph.D.'s seen in Figure 1. If national needs now (also) point in other directions, substantial and predictable federal budget enhancements in those directions can be expected to call forth the same kind of response on the part of young people (or midcareer people) contemplating their careers. However, growth in Ph.D. production without corresponding growth in available jobs for Ph.D.'s may cause more harm than good. There is evi- dence that this has occurred in the past. Anecdotal evidence suggests recent widespread underemployment of some biology specialties, indicating pos- sible "overshooting" too many new Ph.D.'s to satisfy the demand. Goldman and Massy (2001), in particular, argue that funding increases natu-

OCR for page 167
1 2,OOO 1 O,OOO 8,OOO 6,OOO 4,000 2,OOO o Biological Sciences ~ Computer Sciences Chemistry ->I- Engineering, Total Mathematics Social Sciences,Total t ,*,,, Ad, ~~ ~~ ~-w ..... - - - -. . . . . . . , . . I. . . . :: . . . . .-.-.-. . . . . . .. ~ ~ ~ - ' a""' - ' ' ' '''' ~ !!! ~ , ~ ~~......................... 1 1 1 1 1 1 _~ 1 1 '70 '71 '72 '73 '74 '75 '76 '77 '78 '79 '80 '81 '82 '83 '84 '85 '86 '87 '88 '89 '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 FIGURE 8 Federal obligations for total research, by field of science & engineer- ing, FY 1970-2002. Source: NSF Federal Funds for Research and Development Fis- cal Years 1951-2001. rally lead to greater increases in Ph.D. production than in Ph.D. employ- ment, without specific policy interventions. Romer (2000) calls for a system of portable national fellowships as a means of increasing funding for S&E, while allowing market forces a role in matching supply and demand. 2. Increase incentives for private investment and hiring in the priorityfields of science and engineering. This strategy, while less straightforward, falls also in the federal bailiwick. Subsidies, patent and intellectual property protection, and regulatory changes can be effective tools for encourag- ing private investment and jobs in industries that employ particular types of scientists and engineers. Often, jobs fallout is a byproduct of policy intentions toward some other goal, but job growth in particular professions can just as well be the explicit policy target. In either case, young people and others in midcareer can be expected to respond. A1- though not primarily driven by federal policy, the boom in computer science and engineering degrees during the 1990s was fueled by rapidly increasing private sector demand. 3. Adopt the "professional school model" for S&E Ph.D. programs. This strategy aims not at increasing later career rewards but at reducing the early costs and uncertainties of training for an S&E career. The acceptance of this strategy in academe, even any resolve toward attempting it, seems remote. Still, more young people would surely be lured away from pro- fessional schools to S&E doctoral programs, if the years to S&E Ph.D. completion were rolled back, say, to 1970 levels, if this term were predict- able and standard, and if the subjective and arbitrary aspects of the Ph.D. path were curtailed. 4. Introduce two new professional doctoral degree programs for science and engineering, built on the M.D. model. This fourth strategy would also reduce

OCR for page 167
i78 PAN-~CANIZAHONAL SUMMIT training costs and uncertainties, but specifically for those whose career goals focus on professional practice rather than cutting-edge research. Graduates would have a firm grounding in a broad set of skills, under- stand how their skills fit in with other skill sets, and be able to keep up with the cutting edge. These new programs would feature a structured curriculum with well-defined completion criteria and a definite term, per- haps of four years. Their faculty would be practitioners with other sources of income, as in medical schools. The rapid growth of industrial parks, corporate-like technical centers, and corporate partnerships would facili- tate this arrangement. As with existing professional degree programs, stu- dents would not normally rely on grants and fellowships, but would in- stead look to substantially higher lifetime earnings to pay their own way. The attractiveness of this strategy depends partly on whether the current employment of SHE Ph.D.'s could be partly satisfied instead by holders of these new professional doctoral degrees. 5. Expand content and improve timeliness of SHE workforce data. To know whether shortages of scientists and engineers are in fact developing and whether strategies to encourage their production are succeeding, specific additional data should be collected. In addition, a subset of indicators could be developed to provide early warning, some two years or more before full data become available. Logic as well as repeated experience counsels caution in pursuing these strategies, particularly the first two. Young people's career decisions do not shift instantaneously when the relative attractiveness of their vari- ous choices begins to change. Having begun to shift, their choices do not then emerge in the employment market for as long as their graduate train- ing takes and undergraduate training, too, for the many who choose when they are younger. Government actions to raise opportunities and earnings in one field must be sustained for many years or they do more damage than good. To see this, consider that such a policy must be sus- tained for substantially longer than the lag between the policy's initiation and the labor market entry of the last new crop of graduate scientists and engineers. In this last crop are the first high school students who jumped (and were encouraged to jump) in the newly favored direction. Hence, 8 to 10 years is the absolute minimum period of sustained government in- vestment before those young people who responded can begin to reap the reward, much less begin to repay their investment. Policy that cannot be sustained for more than a decade will therefore be destabilizing and harm- ful to bright young people's careers and lives, to the extent that they and their advisers trusted the policy. These important caveats notwithstanding, sustained strategic move- ment in any of these first four directions could reduce the costs and uncer-

OCR for page 167
tainties of postgraduate SHE training and increase the job opportunities, earnings and satisfaction of graduates, whether in priority fields of sci- ence and engineering or across the spectrum. In response, the young people who are bright enough to drive 21st century American science and engineering (but also bright enough to work its clinics, courts, and busi- nesses) would increasingly do so. REFERENCES American Bar Association. (2002) Statistics. Berger, Mark C. (1988~. "Predicted Future Earnings and Choice of College Major." Industrial and Labor Relations Review 41~3~: 418429. Current Population Surveys. (1998-2000~. U.S. Census Bureau. Day, Jennifer C., and Newburger, Eric C. (2002~. "The Big Payoff: Education Attainment and Synthetic Estimates of Work-Life Earnings". U.S. Census Bureau. Federal Funds for Research and Development: Fiscal Years 2000, 2001, and 2002. National Science Foundation, Division of Science Resources Statistics. Federal Funds for Research and Development Detailed Historical Tables: Fiscal years 1951-2001. National Science Foundation, Division of Science Resources Statistics. Goldman, Charles A., and Massey, William F. (2000~. PhD Factory: Training and Employment of Science and Engineering Doctorates in the United States. Bolton, MA: Anchor Publishing. Loftsgaarden, D. O., J. W. Maxwell, et al. (2002~. "2001 Annual Survey of the Mathematical Sciences (Second Report)." Notices of the AMS 49~7~: 803-816. National Research Council (1998~. Trends in the Early Careers of Life Scientists. Washington, DC, National Academy Press. National Science Foundation. Science and Engineering Doctorate Awards (1995-2002). Arling- ton, VA, Division of Science Resources Statistics. National Science Foundation (2002~. Science and Engineering Indicators (2002) Arlington, VA, National Science Board. Romer, P.M. 2000. "Should the Government Subsidize Supply or Demand in the Market for Scientists and Engineers?" NBER Working Paper 7723. Cambridge, MA: National Bu- reau of Economic Research Salary and Employment Survey. (1999) Chemical and Engineering News 77 (31~: 28-39. Salsberg, E. and G.J. Forte (2002~. "Trends in the Physician Workforce, 1980-2000." Health Affairs 21~5~: 165-173.