4
Supply: Who Enters the Profession?

This chapter discusses the sources of talent for computing professional jobs, outlines the broad range of educational programs that specialize in developing skills needed by different kinds of computing professionals, and addresses the need to attract and educate a diverse work force for the computing professions.

INTRODUCTION

Where does the talent come from to fill job openings for computing professionals? Computing professionals are those individuals who have acquired the requisite skills by completing appropriate courses of education and training and/or by gaining appropriate experience. They include a large group of individuals who can fill entry-level positions, which typically have the lowest or most general requirements, and a smaller group of individuals who can fill positions requiring more experience or expertise.

Requirements for education and expertise appear most stringent for research positions, where a Ph.D. is typically expected, at least in academia; they are more variable for applications and systems development positions, where employers hire people with varying degrees and levels of education; and they are least stringent for applications and systems deployment positions, where advanced education is not required (but experience is valuable). Accordingly, the talent pool



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 58
Computing Professionals: Changing Needs for the 1990s 4 Supply: Who Enters the Profession? This chapter discusses the sources of talent for computing professional jobs, outlines the broad range of educational programs that specialize in developing skills needed by different kinds of computing professionals, and addresses the need to attract and educate a diverse work force for the computing professions. INTRODUCTION Where does the talent come from to fill job openings for computing professionals? Computing professionals are those individuals who have acquired the requisite skills by completing appropriate courses of education and training and/or by gaining appropriate experience. They include a large group of individuals who can fill entry-level positions, which typically have the lowest or most general requirements, and a smaller group of individuals who can fill positions requiring more experience or expertise. Requirements for education and expertise appear most stringent for research positions, where a Ph.D. is typically expected, at least in academia; they are more variable for applications and systems development positions, where employers hire people with varying degrees and levels of education; and they are least stringent for applications and systems deployment positions, where advanced education is not required (but experience is valuable). Accordingly, the talent pool

OCR for page 58
Computing Professionals: Changing Needs for the 1990s for computing professional jobs is expected to be the smallest for research positions and the greatest for deployment positions. The willingness of employers in industry to train (and retrain) people expands the available talent pool, especially for development and deployment positions. The willingness of citizens of other countries to immigrate (or to work locally for U.S.-owned employers) also expands supply. Supply is of concern principally as it relates to demand—Are there or will there be enough computing professionals to meet the needs of employers?—and it can be considered from the perspective of current conditions or of anticipated future conditions, which may not be equivalent. Workshop participants seemed to agree that, given the current economic and public policy context, the total supply of computing professionals today is adequate, although in some specific areas (e.g., systems research, systems integration, certain kinds of software development), especially in those with requirements for very specialized applications knowledge or experience, it is not. Representatives from industry differed in their outlook on supply, depending in part on the amount and kinds of hiring they contemplated. One factor that may contribute to an increase in the available talent pool is the slowdown in defense spending, which is expected to free up computing professionals as well as other scientific and technical personnel. Paul Stevens, manager of Corporate Software Initiatives at Hughes Aircraft Company, noted that this effect on human resources may be disproportionate, since commercial projects tend to use fewer people than comparably sized defense projects. Workshop participants seemed to agree that supply may not be adequate as we approach the 21st century, for two reasons. First, as discussed in Chapter 3, skill (and education) requirements appear to be increasing. Leslie Vadasz emphasized this trend as a source of concern and a motivator for action: ''[E]mployees of the future will need more . . . baseline education, . . . higher than it was 10 . . . or 20 . . . or 30 years ago . . . ." Second, too few people appear to be pursuing education and training relevant to work as a computing professional. Betty Vetter underscored the difficulties posed by changing current and future prospects: "One of the worst problems of . . . supply and demand is [seen] in the situation we are in now, . . . where current supply exceeds demand and most of us expect future . . . demand to exceed supply." Vetter also noted that even today, computing professionals are relatively well paid, a sign that they are valued.

OCR for page 58
Computing Professionals: Changing Needs for the 1990s DEGREE PROGRAMS At the core of the talent pool of computing professionals are those individuals who have earned degrees in computer science or a related field. Although in many fields (e.g., physics, chemistry, and various branches of engineering) degree holders constitute a clear majority of those working in the field, in the computing professions a large proportion of employed individuals have degrees in other fields. In part, this unique characteristic of computing professionals reflects the relative newness of degree programs in this field and is typical of emerging disciplines. Over time, as programs in computer science proliferate on campuses, the proportion of individuals who work as computing professionals but have degrees in other fields can be expected to decline. Currently, however, degree production in the field is a weaker indicator of the supply of computing professionals than of the supply of professionals in other fields. Further complicating an assessment of supply based on degree production is the variety of programs offering a degree related to computer science. Before considering the numbers of degrees awarded, it is useful to understand the kinds of degrees being counted. Education and training for computing professional jobs as a group can be gained in 2-year programs (although in general, computing professional jobs require at least a 4-year course of education), baccalaureate programs, master's programs, and doctoral programs. The range of programs is described in a paper (Appendix B) prepared for the workshop by A. Joseph Turner, a professor in the Department of Computer Science at Clemson University, and summarized below. Turner notes that "a program's title in itself doesn't tell you much about what is in the program, and there is a lack of standardization." As Turner explains, attention to degree options (as opposed to separate degree titles) in electrical engineering and possibly other sister disciplines might result in an even greater number and variety of programs than those he lists. Turner's presentation underscores the problems that arise in attempting to categorize and analyze available information on degree production. Two-Year Programs Two-year programs, of which there are several hundred in the United States, are oriented toward producing candidates for entry-level jobs, including people who want to switch fields. They focus more on acquisition of specific skills than on the general principles

OCR for page 58
Computing Professionals: Changing Needs for the 1990s and concepts that characterize 4-year college and university programs. Two-year programs range from combinations of two or three courses in programming languages to more substantial curricula; they may focus on the use of specific applications (e.g., word processing, spreadsheets, or computer-aided design) or on rudimentary programming.1 The scope and rigor of these programs are such that, absent other skills, their graduates are typically considered candidates only for lower-level deployment (e.g., technical support) and technician (as opposed to professional) positions. Baccalaureate Programs Most people who pursue an education in computing complete a baccalaureate program. Baccalaureate programs related to computing, which offer from 6 to 15 courses of varying quality and rigor over 4 years, have a range of names such as computer engineering, computer science and engineering, computer science, computer information systems, information science, information systems, management information systems, and management of information systems. In terms of the types of graduates they aim to produce, baccalaureate programs can be grouped loosely in three categories—liberal arts, professional, and basic science. Liberal arts programs tend to be broad and little oriented to development of specific skills; professional programs tend to focus on development of skills (production of professionals); and basic science programs, much like 4-year programs in the physical sciences, are broader than professional programs and provide a foundation for graduate study. Because of a lack of consistency, the distinction between bachelor of science (B.S.) and bachelor of arts (B.A.) degrees is meaningless in computer-related fields. Recognizing that program and degree titles are not used consistently, it is nevertheless possible to categorize bachelor's programs according to their substantive focus, along a spectrum that includes a hardware (engineering) orientation at one end, a science orientation in the middle, and a management of information systems orientation at the other end. Curriculum guidelines are produced by the Association for Computing Machinery; the Accreditation Board for Engineering and Technology, for engineering degrees; the Computer Science Accreditation Board, for computer science degrees; and, for some 2-and 4-year professional and management-oriented programs, by the Data Processing Management Association (see also "Education Curricula" in Chapter 5). Discussions at the workshop concentrated on skills and curricula

OCR for page 58
Computing Professionals: Changing Needs for the 1990s most closely associated with computer science or engineering; less was said about information systems or information science, in part because of the makeup of the group. While both information systems and information science programs are primarily concerned with computing, they have other emphases. Information systems programs focus on management concerns and on the use of computing and communications systems in the business environment; information science programs focus on the use of technology to organize, store, and retrieve information, with a secondary emphasis on cognitive and decision science. Graduates of these programs are oriented to deployment and some development activities. By contrast, computer science and computer engineering programs are distinguished by their focus on computer science and computer engineering foundations and concepts. Graduates of these programs are oriented to development and, when they pursue further education, to research. Computer science and computer engineering programs overlap substantially, with computer engineering programs putting more emphasis on hardware—although today, many hardware designers tend to come from electrical engineering rather than computer engineering programs. The engineering orientation of computer engineers equips them well for applications work (e.g., in manufacturing). Turner notes that, like other engineering programs, computer engineering programs tend to be well defined. Master's Programs Master's programs produce graduates who are hired for applications and systems development jobs and sometimes industrial research jobs. Turner estimates (Appendix B) that more than 300 departments offer master's degrees in computer science, about 95 in information science, about 72 in information systems, about 33 in management information systems, and about 40 in computer engineering. Master's programs have varying requirements, ranging from course work only to course work plus a project or paper or thesis, and may also offer specialization in such areas as telecommunications, decision support systems, and artificial intelligence. Master's programs can be either traditional science programs or professional programs. In computer science, master's programs have tended to be intermediate steps en route to a Ph.D. They are science and research oriented, providing more education and skills than a bachelor's degree but no particular training relevant to development activity in industry. As a result, and in reaction to feedback from industry, there has been some discussion in the community about the

OCR for page 58
Computing Professionals: Changing Needs for the 1990s need for a terminal professional master's degree in computer science. By contrast, prominent among master's programs are those in the areas of information systems and management of information systems, the latter sometimes related to M.B.A. programs. According to Turner, information systems programs may have significant computing elements (including software development) as well as business elements, whereas management of information systems programs tend to focus more on business and management issues. Two new types of master's programs tend not to be captured in aggregate statistics: master's programs aimed at producing software engineers, of which there are now 15 to 20 in the United States, and master's programs in computational science (the application of computing to large-scale scientific problems), of which there are now a handful. Doctoral Programs Doctoral programs in computer science and computer engineering are tracked in detail in the Taulbee survey (conducted by the Computing Research Association), as discussed in Chapter 2. There are fewer than 200 such programs. Although counting the computer science and computer engineering Ph.D.s they produce is relatively easy, it remains the case that the contents of those programs, and therefore the quality and competence of their graduates, are not consistent.2 Less is known about doctoral programs in the related areas of information science and information systems, which are not covered by the Taulbee survey. Within the academic computer science and engineering community, the optimal rate of production of Ph.D.s has become a subject of debate. Central to that debate are expectations for 1,000 Ph.D.s per year established by the federal High Performance Computing and Communications program, which establishes research and human resources directions for computer science, computer engineering, and computational science. Workshop discussions raised some concern that targets are apparently being based on, at best, imperfect forecasts of demand. As noted in Chapter 3, limited industrial requirements for Ph.D.s and uncertain prospects for research funding create uncertainty about optimal Ph.D. production levels.3 One issue discussed at the workshop was how the mix of ages of people in the field might affect the supply of computing professionals. Betty Vetter noted that as a group, computer science Ph.D.s are considerably younger than engineering, physical sciences, or mathematics Ph.D.s. As a result, computer science will not face the prob-

OCR for page 58
Computing Professionals: Changing Needs for the 1990s lem expected in other fields between 1995 and 2000, when a large number of retirements are expected.4 Vetter cited the American Council on Education's Campus Trends 1991, which observed that its membership, which includes most academic institutions, highlighted computer science as having significant faculty shortages today, but also indicated in response to a follow-up question (''And do you expect this to be better or worse in five years and at about what level do you expect it?") that they expected faculty shortages to be reduced in computer science, and only in computer science.5 The evidence on age distributions and the institutional responses reported by the American Council on Education do not support the hypothesis of future shortages of computer science doctorates as strongly as does comparable evidence for doctorates in other fields. Also, anecdotal evidence from mid-1992 suggests that there is a surplus of candidates for faculty positions, at least for those at upper-tier schools. FUTURE SUPPLY: PIPELINE OR AQUIFER The supply of computing professionals for the future depends primarily on the supply of people capable of and interested in pursuing relevant bachelor's degree programs; the supply of future computer science and engineering researchers depends further on the number of people capable of and interested in pursuing higher degrees, especially Ph.D.s. Adequate supply is a concern even when demand is expected to grow slowly or be flat, both because people will leave the field and have to be replaced and because when there are more qualified people, there will be better odds of achieving a good fit between individuals and jobs. The future supply of computing and other professionals is often discussed in terms of a "pipeline," the numbers of people at earlier stages of education with appropriate interests, aptitudes, and prerequisite course work. Workshop participants expressed discomfort with the pipeline metaphor. They suggested that supply might be described instead in terms of an aquifer. Observed Robert Weatherall, director of placement at the Massachusetts Institute of Technology, We are heading into a desert—we have a lack of a skilled population at many levels. What we need to do is to keep the aquifers filled. . . . Once there are really bright people at the college level and then people getting master's degrees, once . . . the water level is really high, then Ph.D.s will flow. The aquifer model also allows for infusions of talent at different points (e.g., people who switch fields and/or return to school as adults), whereas the pipeline, which focuses on what happens over

OCR for page 58
Computing Professionals: Changing Needs for the 1990s time to children with quantitative aptitude as they proceed through their schooling years, seems to emphasize the process of attrition and a single source for talent. Although workshop participants did not analyze the issue explicitly, a number of comments pointed to the variety of ways in which people have entered computing fields indirectly or at later points in their careers (see, for example, "Experienced Workers," below). Thus, they pointed to the importance of "in-migration" or transfers as a historic source of supply, although their comments about rising skill requirements suggested that opportunities for in-migration might diminish over time or at least involve more formal education or training as formal programs producing computer science graduates continue to diffuse through the higher education system. Encouraging Student Interest Workshop participants expressed strong concern that the number of young people preparing for computing professional jobs is inadequate. To a limited extent, the field may be suffering from a tarnished image: a positive image is fundamental to attracting and retaining talented people in the computing professions. Although these occupations as a group are perceived to be in strong demand and to be well paid, workshop participants suggested that the work of computing professionals may be poorly understood. Moreover, the value of some subspecialties may be underappreciated, even within computer technology firms or academic computer science departments. Participants from industry observed that companies may compartmentalize computing professional activities, limiting apparent advancement opportunities (although some companies offer strong technical-professional career paths). Participants from academia noted that work on large complex systems, which is of special value to industry, typically does not fit well with academic reward structures and advancement opportunities. As a result of these conditions, talented people may choose alternatives to professional computing careers. The problem of attracting students to computing professions was discussed by workshop participants in the context of the more general problems of declining interest in scientific and engineering fields of study and weak preparation in math and science in the kindergarten through high-school period. This point was made by Tora Bikson, who referenced her research on career opportunities:6 [P]eople in corporations, . . . foresee a very short supply of computer scientists and engineers . . . in the United States, especially, . . . 5 to 10 years out, and their perception is that this is so because stu-

OCR for page 58
Computing Professionals: Changing Needs for the 1990s dents aren't being attracted into mathematics and sciences at the grade school and the high school level. Vetter noted that enrollments at the bachelor's level in all of the numerate fields (those requiring a significant background in mathematics) have been dropping since 1986. Brenda Wallace, program analyst at the Bureau of Labor Statistics, described data from the National Center for Education Statistics (NCES) that show that bachelor's degrees in computer science and engineering peaked in 1985–1986 and have declined subsequently. Moreover, the rate of growth and the subsequent decline were greater for bachelor's degrees than for all other degrees in these fields. Participants associated with universities augmented this general observation with comments on their own experiences with declining enrollments in computer science programs over the past couple of years, declines that follow a long period of growth. There is little in the literature that rigorously attempts to disentangle the many factors that might be fueling the decline. Possible reasons for the recent downward trend in computer science degrees include inadequate student appreciation of the rigorous requirements for the degree, expectations that job opportunities are poor, greater attraction of other professional careers, and so on. Declines at the bachelor's level are part of a larger process of attrition, as William Lupton, professor in the Department of Mathematics and Computer Science at Morgan State University, noted in the context of discussing minority interest: There are several leakage points along the way [and] we have to get students through each one of these points. The reason there are so few minority Ph.D.s in . . . computer science is that there are few master's degree holders, and the reason there are few master's degree holders is that there are few baccalaureate degree holders. And it goes right on down the line. While Lupton's statement is appropriate to the production of doctorates in computer science, the concept introduced by Weatherall of an aquifer, rather than a pipeline, allows for the possibility of a flow of talent from other fields—whether into Ph.D. programs or into jobs for computer professionals that require a doctorate. Workshop participants reinforced the notion that interest in computing professional occupations, as in others in science and technology, must begin at an early age, especially for underrepresented groups (Box 4.1). Richard Tapia, professor of mathematical sciences at Rice University, was one of several participants to emphasize the important role of elementary teachers and the need to devote energy to improving

OCR for page 58
Computing Professionals: Changing Needs for the 1990s BOX 4.1 Starting Early "The pipeline has to start long before the student arrives at your front door at the baccalaureate level. If the student hasn't been prepared prior to his arrival at college, he is not going to make it in science and engineering. Therefore, we have to address the pipeline earlier in the student's development . . . [starting] about the fourth or fifth grade. . . . [W]e need to impress upon these students at an early age that science . . . is a good thing to pursue. If we wait until a student has a choice to take a science . . . or a math . . . course and . . . one of his peers says, 'oh, don't take that math course; it is tough . . .,' that plants a seed that will germinate and grow and cause the student to avoid science or math . . . . [W]e need to impress upon students . . . that science . . . is a good thing to pursue . . . before they [decide] . . . that they don't want to pursue math and science. "Further, there is a problem with students getting the proper counseling and guidance, not only from their teachers, their counselors, but . . . from the parents . . . as well. Part of the pipeline issue may be a program to . . . train the elementary school teachers and counselors so that they are sensitive to this and bring the message to the students at an early point, that it is a good thing to do, to tackle that math or that science course. . . . "What we have to do is provide whatever stimuli [are needed] at the early point so that we can maintain these students in the program over the extended period of time. That will require not only training, but monitoring and mentoring as well."—William Lupton their skills: "[E]lementary school teachers . . . love to teach and love kids, but . . . they all have math anxiety. No doubt about it. They don't know this, and they transfer it to their students." Concern was also expressed about inadequate preparation in high schools for college programs requiring significant quantitative achievement or rigor. Although on a national basis relatively few children have significant exposure to computers through their schools, workshop participants challenged conventional wisdom by asserting that precollege programs often appear to discourage rather than encourage interest in computer science and engineering. Although this topic was not discussed at length, possible reasons identified include poorly designed programs and poorly prepared teachers. There was also some speculation that, as computing systems become more common, the technology may lose its allure while the work involved in meeting future requirements may become harder.

OCR for page 58
Computing Professionals: Changing Needs for the 1990s Experienced Workers While typical discussions of supply focus on attracting and educating children and very young adults (candidates for entry-level jobs), older individuals may also provide a source of talent. Although, as a new field, computer science was originally populated only by people trained in other fields, the potential to increase supply today by retraining is often undervalued, according to some workshop participants. Paula Stephan, for example, spoke of the changing demographics of college students: [O]ne of the great paradoxes in recent years in education has been why we continue to have more students in colleges and universities, despite the baby bust generation coming of age. The answer to that is that . . . the huge growth in the United States [has been in] students of non-traditional age. That means that [many of] the people who are knocking at the door for higher education are in their late twenties and their mid-thirties. . . . Tora Bikson reinforced Stephan's comments by drawing on her interviews with employers and professionals. She has found that increasing numbers of people want to be able to move between work and education, although many universities do not encourage this: It used to be that there were two kinds of students, those who get a degree and go to work and those who get a degree and then go into graduate school . . . [but] there [are] more and more cases of students who would like to work for a while after their baccalaureate degree, to get a sense of what the career perspective looks like, sometimes just to earn money for a change or to escape the pressures of school and into the pressures of work. But very many times the traditional faculty doesn't look positively on someone who has polluted [his or her] career this way. Workshop participants from industry provided another perspective on late entrants to the field in discussing efforts to shift personnel from one type of work to another. Leslie Vadasz observed that corporate restructuring is changing demand in ways that will be difficult to interpret or even to capture in data: Corporate America is under tremendous pressure to downsize, restructure. . . . [This situation] is going to create all kinds of misleading . . . data for anybody who is looking at a pipeline, because . . . [they] won't see demand. Yet, at the same time, demand is there in a different way . . . in retraining your own people or in terms of . . . new skills. . . . According to John McSorley, Apple finds it can retrain only about 30 percent of the people whose positions it eliminates, turning out-

OCR for page 58
Computing Professionals: Changing Needs for the 1990s side for new people to fill the majority of its openings. Elizabeth Nichols noted that IBM has had success in training mechanical engineers, among others, for computing professional positions. One reason for historically limited levels of retraining in industry has been the relatively high levels of mobility among computing professionals. Commenting on the West Coast labor market, for example, Paul Stevens noted that computing professionals tend to be more mobile than engineers. McSorley suggested that today, software engineers continue to be particularly mobile, as evidenced by an attrition rate four times that of hardware engineers at Apple. William Gear suggested that the mobility of U.S. computer scientists may be a potential area of American vulnerability: I view this mobility issue as perhaps one of the most serious problems facing American industry. I see there too [much] mobility, . . . causing industry to have little incentive to train people. The emphasis on the individual in American culture, particularly in research, I think is one of the things that perhaps puts us behind the competition in [Japan]. . . . An American researcher sees his or her resume as basically the only employment card for the future. In Japan, employees see service to the company as their guarantee to the future. Marvin Zelkowitz echoed McSorley and Gear in noting that mobility is high for some specialties: ''In computer sciences, . . . good people in more practical areas are really hard to find at almost any level, and there is extreme mobility among artificial intelligence (AI) systems, software engineering, and that type faculty, mostly out of the university system." Even in industry, good people may be lost as technical resources in instances where rewards are perceived as greater in such nontechnical jobs as management. Accordingly, some companies have experimented for years with dual career paths, although the greatest success tends to be within the computer industry; major users (e.g., banks or insurance companies) appear to have more difficulty establishing technical career paths that offer sufficiently rewarding upper levels. PROMOTING DIVERSITY The total supply of computing professionals is only one part of the picture; another important part is the composition of the supply in terms of how well different sociodemographic groups are represented. Noting that enrollments for women appear to be dropping

OCR for page 58
Computing Professionals: Changing Needs for the 1990s faster than the average and remain very low for non-Asian minorities, Betty Vetter summed up multiple data sets as follows: There are a lot of major differences in these data, but all the data show three general trends . . . an increasing and quite high proportion of foreign nationals, . . . small and not increasing numbers and percentages of women, and an infinitesimal number of African Americans and Hispanics. Asians . . . seem to be either appropriately represented relative to their 3 percent of the population or overrepresented . . . .7 Equality of Opportunity Discussions at the workshop underscored the concern about equality of opportunity and breadth of participation in computing professional occupations, both of which are considered essential for the health and growth of the field. Paul Young, associate dean of engineering at the University of Washington, set the stage in his overview comments for the discussion of these issues: I think we all understand that . . . if the United States wants to maintain a good technical work force, changing demographics will force us to have more women and more minorities in technical positions. . . . But aside from the issue of the demographic needs . . . science and engineering are areas that give economic advancement to people, and what is currently happening is excluding a major part of the American population. . . . The record of computer science, while it looked good a few years ago, is, if anything, getting worse, at least with respect to women, and it has never been good with respect to underrepresented minorities. . . . One of the things I was struck by as I listened to members of our panel . . . is that some of the factors that seem to keep women and minorities from computer science probably also account for the general decline in interest in computer science across all populations. . . . Richard Tapia reflected on his experiences in encouraging Hispanic student interest and emphasized that the true problem is one of female and minority participation in science and engineering as a whole: The problem we are talking about is certainly not unique to computer science. In mathematical sciences, [there are] 1 percent Hispanics and African Americans. So, it is really [a problem for all of] science and engineering. And the problem should be attacked from that point of view, not that this is something that computer science has failed on. . . .

OCR for page 58
Computing Professionals: Changing Needs for the 1990s [T]he population of women in the United States stays constant, [while] representation in science and engineering [as a whole] is slowly going up. With minorities, the population is growing extremely rapidly, and yet, their representation is at best staying constant . . . . Of course, I would not argue that it is [a crisis] for the profession—I think we have seen a history in mathematics and computer science that we import solutions. We don't have to turn to our local population. On the other hand, I think it is extremely important that we realize that this extremely large, expanding population is outside of all science and technology, and that this is bad for the country. Workshop participants were concerned that current labor market conditions undercut shortage as a major motivation for encouraging broader participation, although they expected that scarcity may be a problem in the future. It should be noted, however, that even a stable population allows for changes in the mix of participants—even without growth, increasing diversity can be achieved without actively taking jobs away from white males. This is so because some amount of voluntary turnover (e.g., retirements and resignations) is normal, and the vacancies that result provide opportunities to increase representation from previously underrepresented groups. But workshop participants did not foresee a steady-state situation; rather, they anticipated slow growth, and growth can increase the number of opportunities available for all kinds of candidates. In addition to the growth argued for, there are other compelling reasons for efforts to improve the participation of underrepresented groups. Argued Young, "You can't have groups that are permanently shut out, and you can't get real participation . . . until you get enough people in. It seems to me that we have to continue to make that argument on . . . moral grounds." There are also arguments for promoting diversity that are based on doing good science. For example, women and underrepresented minorities bring to science perspectives that push research efforts in directions in which they might not otherwise go.8 Paula Stephan noted the political dimension: "If you think about the importance of science to our culture and . . . to growth in the U.S. economy, the idea of evolving into the 21st century and being a society in which certain groups are almost virtually [without] representation . . . creates enormous political problems . . . ." A final argument for promoting diversity is based on the under-employment of talent implied by the low levels of participation of these groups. This underemployment represents a cost to society that could be reduced by altering the demographic mix of computing

OCR for page 58
Computing Professionals: Changing Needs for the 1990s professionals, even if total employment remains the same. Economists attribute this cost to market failure arising from the effects of discriminatory behavior.9 And Robert Weatherall characterized the need for broader participation as economic: [I]t is a question of the skills infrastructure. Compared to Japan, . . . Germany, . . . and Singapore, this country has a lack of skills across the population at different levels . . . . [E]ducation disciplines are the vehicle for [providing general] skills—mathematical skills, writing skills, social skills . . . and [learning] computer science is a good . . . educational vehicle. So is mathematics. So is physics. . . . [A]ll [should] contribute to the skills infrastructure. Minorities The problem of limited participation in computing professional jobs is particularly stark for African Americans, who typically account for zero to two computer science or engineering Ph.D.s awarded per year, with eight the annual maximum ever recorded in the Taulbee survey (in 1990–1991).10 Throughout the period from 1970 to 1991, these numbers translated to either 0 percent or 1 percent of computer science Ph.D.s awarded each year. Correspondingly, a negligible number of African American faculty work at Ph.D.-granting institutions covered by the Taulbee survey. William Lupton, also then president of the Association of Departments of Computer Science and Computer Engineering at Minority Institutions, noted that undergraduate interest in computer science and engineering has remained relatively high in colleges and universities historically attended by African Americans, although there, too, there was a surge in the mid-1980s. Alan Fechter speculated that some of this numerical strength may have come from a reorganization and renaming of programs previously associated with mathematics or other sister disciplines and since designated as computer science and engineering programs. Peter Freeman noted that the 1980s also saw increased activities by major companies to stimulate interest among and provide financial support for minority youth: AT&T, for example, and IBM have very strong programs . . . providing scholarships, . . . making regular visits to those campuses, sending people, sending equipment, and then guaranteeing jobs when those people get out. . . . [T]he really bright ones should be going on in the graduate pipeline, and we are seeing very explicit evidence that they are so attracted by the salary in going to work at Bell Labs, Bellcore, or IBM that they don't go on to graduate school. . . . But . . . to me, that is [also] a positive outcome . . . .

OCR for page 58
Computing Professionals: Changing Needs for the 1990s Lupton echoed Freeman's ambivalence, noting that some corporate recruitment programs also provide support for further education. Joseph Turner pointed out that diversion of talented minority students from Ph.D. programs translates into a smaller pool of candidates for faculty positions and a virtual absence of role models for future minority students. Richard Tapia noted that underrepresented groups fall into two categories: obviously talented women and minority graduate students are courted and may have a choice of sources of financial support or good jobs in industry, whereas the larger pool of potential talent remains undeveloped. Wade Ellis, professor in the Mathematics Department at West Valley College, noted that the problem of the larger pool of minority students is compounded by the fact that people of limited means tend to go to schools with limited resources. The limited participation of Hispanics in computing programs presents some of the same challenges as participation of African Americans. Within Ph.D.-granting institutions covered by the Taulbee survey, Hispanics compose 1 to 2 percent of Ph.D.s produced and faculty employed in computer science and computer engineering. Understanding the Hispanic situation is complicated, Tapia explained, by the fact that published data tend to overlook the heterogeneity of Hispanics. Although over half of U.S. Hispanics consist of Mexican Americans and Puerto Ricans, most of the people counted as Hispanics in science and engineering appear to be from Central America—a minority within a minority—and the majority of Hispanics have negligible participation in science and engineering. Women In Ph.D. programs in computing, the overall representation of women is low when judged by the fact that they represent one-half of the talent pool, but high when judged by their representation in other fields of physical science or engineering: women tend to receive less than 15 percent of computer science and engineering Ph.D.s, having earned 12 percent in 1990–1991 in the Taulbee-surveyed institutions;11 other sources of data cited by Betty Vetter show the same order of magnitude. Among Ph.D.-granting programs covered by the Taulbee survey, women compose 8 percent of computer science faculty and 4 percent of computer engineering faculty totals, including only 4 to 5 percent of full professors. Although women share the problem of underrepresentation with minorities, increasing the participation of women overall is expected to require a somewhat different set of solutions than increasing participation of minorities.

OCR for page 58
Computing Professionals: Changing Needs for the 1990s Among minority students, women appear to outperform men. According to Vetter, Female African American students do far better relative to African American males, than do female students in any other racial or ethnic group relative to the men within that same group. In engineering, for example, women make up 30 percent of the engineering baccalaureates awarded to African Americans and only 15 percent, approximately, in all the other fields. But in computer science, the African Americans have [had] a really big climb in the last years . . . . Tapia indicated that among Hispanics, women also sought computing and mathematics degrees more actively than men. However, Vetter pointed out that the prominence of foreign students in science and engineering programs tends to work against growth in the participation of women: ''The inclusion of more and more foreign students is detrimental to the [measured participation] of women . . . in any field at the Ph.D. level . . . because foreign students will include far fewer women than American students will." Vetter also noted that language and cultural differences between foreign and American students appear to discourage American students, who often have difficulty with foreign-born teaching assistants. While there is substantial anecdotal evidence of corporate programs to stimulate the interest of women and minorities in science and engineering, Tapia contended that Ph.D.-producing universities do not present as welcoming an environment, something evidenced by the paucity of women and minority faculty in those institutions. Lucy Suchman reinforced the concern about the environment in Ph.D. programs, which she suggested is particularly hostile toward women as "polluting" elements. [W]hat actually happens to people when they encounter the institutional realities . . . ? That is what makes people drop out and that is what makes the barriers higher as you go up. For women, they come in great numbers and as we look up the management hierarchy, they get harder and harder to find . . . . [Female graduate students in computer science] . . . get demoralized in absolutely astounding ways. . . . [T]he conditions . . . they encounter [are] . . . those of a very tightly knit club . . . which they are polluting. As a consequence, they suddenly find themselves extremely marginalized and taking it all on as a personal shortcoming, rather than as a limit of the situation. Paul Maritz, as an employer, noted the difficulties he faced as a recruiter in search of women and minority candidates at top-tier schools. Robert Kraut observed that one way to involve more women is to

OCR for page 58
Computing Professionals: Changing Needs for the 1990s redefine the field in a broader manner, to embrace the multidisciplinary nature of some of the current challenges. Suchman echoed this argument, suggesting that new models should be considered. Computer science in Denmark is a . . . very formal discipline, . . . traditional, . . . narrowly defined, but [increasingly] there are . . . departments of information that are co-located with departments of computer science, and they have very high representation of women. Basically, what they are about is a much more complex assembly of activities, including a lot of attention to the intrinsic requirements of the technology . . . . Those kinds of programs . . . really redefine what the whole enterprise is about in a way that, at least for the women . . . I have talked to, . . . [makes it] a much more attractive field to get involved in. Paul Young revisited related concerns in reporting on his conversation during the workshop with Essie Lev, counselor, Information and Computer Science at the University of California at Irvine (UCI). Based on her experience with UCI's computer science department, Lev said, "One of the reasons we are losing students in general, and women and minorities in particular, is that we have lost a sense of passion and a sense of social relevance about what we do." Historically, women (and perhaps to a lesser extent minorities) have sought out fields, such as biology and the social sciences, that emphasize social relevance and teamwork. By emphasizing those areas of computer science and computer engineering that most require those qualities, it may be possible to attract more women (and minorities) to computing professions. FOREIGN-BORN CITIZENS Workshop participants indicated that an important question is whether the pipeline for computing professionals is becoming global. Although they recommended that the question be studied more systematically, indirectly they appeared to answer the question affirmatively. Computing professionals of foreign citizenship appear to be a significant presence within the United States. They are key elements in staffing some research and educational institutions and are prominent among student populations, especially at the graduate schools. For example, NRC data indicate that 54 percent of computer science Ph.D.s from U.S. universities in 1991 were earned by foreign citizens.12 Moreover, foreign students are often better prepared than American students. A large number of the foreign-born students stay

OCR for page 58
Computing Professionals: Changing Needs for the 1990s in the United States after they complete their educations and become part of the domestic talent pool. At the same time, the growing stock of foreign computing professionals and the globalization of various markets mean that computing professionals located abroad are becoming more attractive to U.S. employers, as discussed in Chapter 3. Moreover, advances in communications systems and cheaper communications facilitate the use of remote staff, whose contributions can be transmitted quickly and easily. Immigration may also provide a significant amount of talent to our domestic pool. The contributions of foreign-born scientific and technical personnel to the country have been recognized in recent immigration reforms, intended to facilitate recruitment and retention of foreign scientists and engineers in the United States. ISSUES AND CONCLUSIONS Greater and more effective efforts are needed to attract a more diverse group of people to computing professions to more closely reflect the demographic makeup of the country and assure survival of the field. In particular, women and non-Asian minorities ought to be encouraged to choose and continue in professions in computing fields. Although differences in educational curricula across institutions are consistent with the rich and dynamic nature of the field, an excessive degree of variation may be counterproductive. For those who are interested in a computing career, the educational arena presents an array of choices so broad as to be confusing. Virtually every dimension, from labeling of programs to content, seems to vary, in every class of program. Perhaps improvements in taxonomies for data collection (see Chapter 2) can serve as a guide to educational institutions and students, helping to focus the skill development process. While computing professional jobs are likely to continue to attract people from other fields, increasing skill requirements may call for more fundamental education in computer science and engineering than was needed in the past. However, the trend toward greater cross-disciplinary interaction in addressing problems in computing and other domains might serve to fuel further in-migration. How different trends will balance out is unclear at this time. Because skill needs are changing rapidly in industry, educational institutions need to produce graduates with the ability for and commitment to continuing future learning. In theory, these are

OCR for page 58
Computing Professionals: Changing Needs for the 1990s consequences of good programs of basic education; in practice, it is not clear that these objectives are being met. NOTES 1.   A set of new curriculum recommendations is under development by a committee of the Association for Computing Machinery, a professional society. This group has classified 2-year programs into five categories, four of which are relevant here, including computing and engineering technology (hardware-oriented programs), computing and information processing (information systems-type courses), computing science (more computer science oriented), and computer support services (including functions like computer operation). 2.   This problem is not unique to computer science and computer engineering, but many have the impression that the situation is more serious in computer science and computer engineering. 3.   See Computer Science and Telecommunications Board, Computing the Future: A Broader Agenda for Computer Science and Engineering, National Academy Press, Washington, D.C., 1992, for a discussion of research trends and needs. As that report notes, the opportunities for Ph.D.s will be greater where individuals have broader views of what constitutes interesting and legitimate research activity. Although employers may specify a requirement for Ph.D.-level education relatively infrequently, Ph.D. holders have skills that could be applied beneficially in nonresearch activities, including development. However, as Elizabeth Nichols observed, employers sometimes have difficulty merging Ph.D. holders in development roles. 4.   Retirements and deaths are increasing in Ph.D.-granting institutions. The Taulbee survey reported 35 such separations in 1990–1991, more than twice the level of the previous year, a level that had also exceeded that of earlier years. 5.   Elaine El-Khawas, Campus Trends, American Council on Education (ACE) Reports, Number 81, ACE, Washington, D.C., July, 1991, pp. 10–12, 36. The statistics refer to all institutions of higher education, including community colleges. If the analysis is restricted to doctorate-granting institutions, an increasing proportion of the respondents expect future shortages. This finding applies to each of the individual fields covered in this survey, including computer science. 6.   T.K. Bikson and S.L. Law, Meeting the Human Resource Needs for Success in a Global Economy, College Placement Council, Bethlehem, Pa., 1992. 7.   Vetter noted that the discrepancies in statistics about Asians suggest that some surveys may count Asian Americans with foreign nationals, and some, appropriately, may not. 8.   Evelyn R. Keller, Reflections on Gender and Science, Yale University Press, New Haven, Conn., 1990. 9.   See, for example, Gary Becker, The Economics of Discrimination, University of Chicago Press, Chicago, 1971. 10.   David Gries and Dorothy Marsh, "The 1990–91 Taulbee Survey Report: The Computing Research Association's Survey on the Production and Employment of Ph.D.s and Faculty in Computer Science and Engineering," Department of Computer Science, Cornell University, Ithaca, N.Y., December 1991. 11.   These numbers include women of multiple nationalities, not just U.S. citizens. 12.   Data from annual and other surveys of the National Research Council's Office of Scientific and Engineering Personnel.