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.
CHAPTER 1 FACTORS AFFECTING THE SUPPLY OF DOCTORAL WOMEN SCIENTISTS AND ENGINEERS Women have traditionally constituted a small minority of all doctoral scientists and engineers, although their representation varies considerably by field across the spectrum of scientific and technologi- cal disciplines. During the decade of the 1970s, for example, women earned about a quarter of all social science doctorates and one-fifth of those in life sciences but less than one-tenth in physical sciences and only 2 percent in engineering. The proportion of women in all these fields has been rising, dramatically in some areas, but despite this trend women's relatively low apparent interest in scie.noe fields remains a central issue in any assessment of women's status in the sciences. Although the dominant concern of this rep-~r.t As to examine the progress in academic careers of those women who do choose to pursue science, it is necessary also to explore the factors that may affect such a choice for good or ill throughout the course of prep-ara- tion. In this chapter we will examine both individual and structural factors that contribute to educational and career choices. Historical patterns and their reflections . In a general framework of human resources studies there are two basic preconditions for achievement in science or any other calling: ability and opportunity. The interactions between these two factors can lead to many different results, but two boundary conditions exist in the real world: 1) even outstanding ability needs to be trained, directed, provided with a sphere of action, and rewarded in Order to come to fruition--in short, to have opportunity to flourish; and 2) ~ 1 ability does not exist. Historically, this quite obvious generalization has been seen as valid for men, leading especially in the sciences -in the period following World War II to the creation of an elaborate system of educational and career support opportunities (pre- and postdoctoral fellowships, career service awards, research support, eta ~ predicated on two assumptions: 1) that whatever scientific talent exists initially must be nurtured by enhancing appropriate opportunities, and 2) that it is in the national interest to do so. even unlimited o~oortunitv cannot create achievement where reaui.~te 1.1
In the case of women's careers in the sciences, however, these simple assumptions have not been applied in the same way, with the result that both women's access to training in science and the later development of their careers have been forced into a different mold. From the mid-19th century when women began to be admitted to higher education up to the late 1960s, the dominant rationale for educating women at all was that it would make them better wives for educated men, and better mothers for their children. True professional career preparation, especially in the sciences, became a reality only for a handful of outstanding women in each generation. Some of the factors that still affect adversely women's access to and development in science careers are directly traceable to these limited perceptions. Aside from the issues of potential sex differences in cognitive ability discussed more fully below, women's aspirations and achievement motivation have been examined and questioned from a perspective that assumed the absolute primacy of women's "instinctive" desire to marry and raise children without evaluating how realistic their career-related motives and aspirations might indeed be in light of the actual oppor- tunities open to them. In contrast to this position, the Committee makes only one assumption in examining the factors that affect women's achievement in science: that all individuals, male or female, make educational and career choices in a way that they believe will maximize their chances for a productive, rewarding, and happy life within the frame- work of opportunities they perceive as real. Such an assumption has the advantage of enabling us to interpret rationally the recent trends of steeply increasing participation of women in science as a normal and expected response to expanding opportunities, rather than having to postulate a dramatic, fundamental shift in female psychology occurring in the 1970s. In the remainder of this chapter we examine evidence relating to cognitive ability, opportunities for access to science careers, and how the interactions between them may affect men and women differently. We use a "pipeline/valve" model of the education-career sequence as a conceptual framework for this examination. In such a generalized model, the valves represent successive selection processes in attainment, such as represent successive selection processes in attainment, such as undergraduate and graduate degrees, junior faculty appointments, and promotion to tenure. Among other properties, such a system has the characteristic that if the final valve in the line is constricted even if all others remain open, the flow will decrease. Flow will be dependent on both external factors such as availability of places in graduate school, financial support for students, and professorial appointments, and on internal or self-selection of individuals who assess by whatever means and information accessible to them how good their chances are of reaching a desired stage. Women's decisions to pursue or abandon careers in science will thus be evaluated against the background of real or perceived opportunities open to them at several periods in time. Note that perceived opportunities are equally 1.2
if not more important than real ones, since individuals make choices based on their perception of a situation, which may differ from the facts. Scientific ability The assumption that women's general intellectual endowment is inferior to men's is an ancient one which was elaborated during the 19th century by such diverse disciplines as psychology (Shields, 1975) and physical anthropology (Gould, 1981) and publicly proclaimed on innumerable occasions during the debates surrounding women's admission to higher education (Conable, 1977; Earnest, 1953~. More recently, this assumption has become less sweeping, restricted primarily to putative sex differences in mathematical ability; if women's general mathematical ability or some relevant segment of it can be shown to be inferior to men's, such a finding would explain low participation in science, especially in the most quantitative areas. It has been suggested that differences in mathematical ability may correlate with differences in brain lateralization (McGee, 1979), and most recently in androgen levels (Hier and Crowley, 1982~. In the latter case, low spatial ability and low androgen levels were shown to coincide in males, although no similar quantitative correlations were attempted for females. It is perhaps too early to tell whether such conjectures are as erroneous as the 19th century focus on sex differences in brain size, which is now recognized to be related to body size (Gould, 1981) rather than intelligence. Kagan (1982) has pointed out a number of fallacious assumptions in Hier and Crowley's work. The information in the literature that bears on sex differences in scientific ability is restricted largely to aspects of mathematical ability, which are used as a proxy measure. The connections between such discrete factors as spatial visualization, abstract reasoning, or analytical ability (which have been tested) and successful scientific achievement seem intuitively obvious to most scientists, but the nature of the relationships have not been analyzed and tested. In broad outline, well-defined sex differences are found in scores on large-scale standardized tests such as the Scholastic Aptitude Test - Math (SAT-M) and the National Assessment of Educational Progress (NAEP), with girls on the average scoring lower than boys on the SAT-M by about 0~4 standard deviations. Most but not all of the difference disappears when the results are controlled for the number of math courses taken in high school and for whether or not individuals are enrolled in a math course at the time of the test; the test therefore almost certainly represents a measure of achievement rather than innate aptitude. When test results are properly controlled for course-taking, a small difference favoring males remains in segments testing problem-solving for those students who have not taken calculus or geometry but no difference is found between boys and girls who have
taken these courses (Armstrong, 1979). A number of studies (reviewed by Lockheed, 1982) have examined the possibilities of sex bias inherent in a variety of standardized tests, primarily in terms of either content bias or psychometric bias. Such types of bias have not been demonstrated conclusively, and Lockheed con- cludes that the differences in scores represent real sex differentiation · · . an learning experience. Sex differences in either mathematical performance or interest are not observed in young children but differences in expressed interest appear among academically highly able students in 7th or 8th grade, with boys much more interested in math. This greater interest appears to be closely linked to career interests in scientific fields (Haven, 1972~. Other findings (Astir, 1968; Fox, 1981; Helson, 1971) support the conclusion that timely encouragement during adolescence is important in sustaining mathematics-related interests of girls and boys, but that girls are far less likely to receive it from any source. In contrast to findings of lower mean scores for women on national tests, other researchers (Fennema and Sherman, 1977) find that in certain schools or classes no sex differences in mathematics achieve- ment can be demonstrated concluding that this outcome results from absence of various kinds of sex bias in such schools. In any event it is clear that from the point at which math and science courses become elective, girls have in the past chosen (or were allowed) to participate less than boys, resulting in female high school graduates having had about one-half course less than males in these fields (National Center of Education Statistics, 1979~; to what extent the courses girls do take may differ from those taken by boys (e.g., business math versus algebra II) is not clear. In the aggregate, sex differences in high school science and math preparation or in existing aptitude measures appear to be too small to account readily for the disparate distributions of women and men in these fields in college. For example, 46.4 percent of male high school students versus 43.1 percent of females take college preparatory math (including algebra II or III, geometry, trigonometry, pre-calculus, and calculus) and 21.8 percent of boys versus 17.9 percent of girls take chemistry or physics (NCES, 1981~. Precise information on the high school sex distributions in calculus or advanced chemistry and physics courses is not available. One qualitative study of high school girls in Advanced Placement math and science courses (Casserly, 1979) reports pervasive efforts by counselors to steer girls out of these courses, chiefly on grounds that they would have to work too hard or might "spoil a good record." Male counselors who tried to dissuade girls from taking such courses are quoted as worrying that these girls might take needed jobs away from men. One director of guidance stated (p. 12~: "There are men with Ph.D.s in physics all over the place who can't get jobs. Why 1.4 1
should we encourage girls? Why, if they're successful, they'd be taking jobs away from men who need them. No, it wouldn't be fair to the girls." Nonetheless, several studies in the late seventies suggest that sex differences in course taking are disappearing (cited in Fox, 1980, p. 8) as a result of increased participation by girls. Fox concludes that these changes are still in progress but are occurring faster in some age, ability, and/or geographic groups which may account for variation in findings. Such changes should be reflected in increased participation at the college level within the next two or three years and by the late eighties at the Ph.D. level. Trends in math and science ~reparation_~n college Student distributions by sex among various college majors are customarily measured as percent of baccalaureate degrees earned, a misleading indicator which does not take account of the changing sex composition of the total student body over time. For example, between 1970 and 1980 the proportion of total baccalaureates earned by women rose from 43.2 percent to 49 percent, with two-thirds of the increase occurring in the second half of the decade. The growth in women's participation in science is measured more usefully as the ratio of the percent degrees earned in a given science field to the percent of total baccalaureates earned by women in a given year, as in Table 1.1. This ratio also has the characteristic that it approaches unity as the sex distributions converge, and thus constitutes an accurate measure, termed "parity index," of the relative interests of college men and women in a given field at a particular time.) Parity indices for female participation in selected broad fields of college science are given in Table 1.1. Steeply increasing indices are noted in those fields in which women have been most underrepresented, i.e., engineering and computer sciences. Mathematics, long the most sex-neutral of college majors, shows a very slight total decline in popularity among women over the last decade after peaking in the early 70s, while during the same period women's relative participation in physical sciences has increased by about 55 percent. In biological sciences, historically the most popular of the natural sciences among women, participation over the decade has risen by about one-third and is now approaching the male distribution. iThe prior question posed by using this measure--the determinants for both sexes of choosing to attend college at all, and choosing which college to attend--is beyond the scope of the present report but is addressed fully in Perun, 1982. Significant sex differences in finan- cial support generated by the G.I. Bill in the post-World War II era account to a large extent for sex differences in college attendance. 1.5
. - a) o o 1 o ~9 u] .- a) v u) . - o 3 o U) a) v v Q ,1 3 o . r.1 - m in; En H Pi m of . - ·,1 To . - U] H A, O m o . - o\o m . - ·,1 Pi H m To v H m ~0 On ~ m 0 ~ 3 o To ~Ln~ ··~ ED ~ ··~ o CO o . Go [Ior o ·· ~ ~9 ·~ co<9 ~D ~[ ~0<9 ~k9 ~oo ····ce·eee ~O ~00 ~D;`~ ·· ~··e··ea~ CO03 ~03 C~ . ~r[ ~<9oo d. ~ e ··e··~ L ~00Lr) ~(9 ~U ~ ····e··e c ~D ~ 03COO ~O ~00\9 [~[~CO00 ~00)00~1000000 ·e······e ~ ~0 ~r0 ~r ~cn ·····e··ee [~431 ~OO rnrn ~r ~ O ~O ~d1U] ~O ······a·e·e U ~OO ~U~[~ ~m) ~4 _ . O ~O ~O ~Dv9<9~)k9t~ 1. 6 Q -- ~ ~-- O .- U] a) -,' U) ~; m ~0 o U] 0 ~ a -,' -,1 ~; m 0 3 dP a Q~ 11 O m H C) o ~ ~Q u 11 U] ,{ V? .~ U] o . - o o O 3 o ~0 ,<C m 0 dF ·e rr1 ~; o ~Q
Table l.lA, showing trends in women's participation in English, foreign languages, and education majors is included for comparison with the trends in science fields. The proportional increases in the popu- larity of science fields among women are comparable to the decreases in these non-science fields; however, it should also be noted that the numbers involved are far larger in education (nearly seven times as large as physical sciences) and English (about twice as large). The numbers of foreign language majors are about the same as mathematics. These opposite trends in field distributions for women obviously do not reflect a simple trade-off, but rather some very broad shifts in both expanding options and interests of women students. These measures demonstrate a real growth of interest in science fields among women beyond the increases deriving simply from the fact that the proportion of total baccalaureates earned by women has grown by about 13 percent over the decade. In terms of the pipeline model, some previously constricted valves have been opened. The causes of this real increase have not yet been addressed in focused research studies. One possible causal factor would be better high school preparation of women students stimulated by a variety of career information and work- shop programs during the 70s, many of them supported by the National Sci- ence Foundation but now eliminated on budgetary grounds. Heightened aspirations and career expectations fostered by the perception of sex equality generated in the last fifteen years can be assumed also to be an important contributing factor; the expectation that women's career rewards in the sciences will approximate or equal men's did not realistically exist much before 1970. Certainly the near equalization of financial support for both sexes that has occurred as a consequence of equal rights statutes (specifically Title IX of the 1972 Higher Education Act and the Equal Credit Act of 1974) has made a significant contribution to equalizing college attendance and possibly has prompted women to choose more often those fields in which graduate work is also likely to be necessary for full professional development. Trends in graduate education The decision to undertake graduate training represents an important valve in our pipeline model; if men and women base these decisions on different information or if the same information has different signifi- cance for the two sexes, then the decisions can also be expected to differ. In the preceding section we have examined how women's initial undergraduate field choices have changed in the last two decades to approach the male patterns of field distribution more closely. In this section we will examine to what extent women's decisions to pursue Ph.D.s in the sciences exhibit a parallel trend. To assess how much variation needs to be explained, it is instructive to look at both persistence/ attrition rates and at a measure somewhat analogous to the parity indices described in the previous section. 1.7
Table 1. 1A Trends in Proportions of Baccalaureate Degrees Earned by Women in Selected Humanities Fields and Educate on, 1960-1980 Women as Percent English Foreign Lang. Education Year of all BAs Percent BA PI Percent BA PI Percent BA PI ~_ 1960 35 e 0 62 ~ 3 1 ~ 78 65 ~ 8 1 ~ 88 71 ~ 1 ~ ~ 03 1962 37 ~ 4 64 ~ 9 1 ~ 74 68 ~ 6 ~ ~ 83 73 ~ 2 1.96 1964 40~0 66~4 1~66 72~7 1~82 76~2 l.S1 1966 40 ~ 3 66 ~ 2 1 ~ 64 73 ~ 0 1 ~ 81 75 ~ ~1 ~ 87 1968 4104 67 ~ 3 1 ~ 63 74 ~ 6 1 ~ 80 75 ~ 9 1 ~ 83 1970 4302 66~9 1~55 74~7 1~73 75~0 1~74 1972 4307 65 ~ 8 1 ~ 51 75 ~ 5 1 ~ 73 74 ~ 1 1 ~ 70 1974 44 ~ ~63 ~ 9 1 ~ 44 76 ~ 6 1e 73 73 ~ 5 1 ~ 66 1976 45~5 62~6 1~38 76~8 1~69 72~8 1~60 1978 47. 1 63 ~ 6 1 ~ 35 76 ~ 4 1 ~ 62 72 ~ S 1 ~ 54 1980 49 ~ 0 ~~ 75 ~ 9 1.55 73 ~ 2 1.49 a. PI = Women BAs in f ield ~ o f all BAs to Women Source: Compiled from National Center for Education Statistics data. 1.8
TABLE 1.2 Persistence and attrition of women from the science and engineering educational ladder ~ Women among: Entering college freshmen, Bachelor's Master's Doctoral by probable degrees degrees degrees majora awardedb awardedb awardedC 1968-69 1971-72 1973-74 1978-79 Mathematics/statistics 45.9 39.1 31.0 15.4 Computer sciences --- 13. 6 12.9 12.9 Physical sciences 14 .7 15.1 14.6 11. 5 Physics --- 6.9 8.2 6.6 Chemistry --- 19. 4 21.8 14.0 Engineering 1.3 1.0 2.3 2. 5 Agriculture 2.0 5.5 9. 8 9. O Biological sciences 36.7 29.6 30.6 26.8 Social sciences 63. 3 38.5 31.8 33.0 a Derived from statistics published in National Norms for Entering College Freshmen, Fall 1968, American Council on Education. As reported in the series Earned Degrees Conferred, 1~.V.~. Education Statistics. ;' Center for Summary Report 1979: Doctorate Recipients from U.S. Universities, National Research Council. 1.9
Persistence and attrition The perception that women are far less likely than men either to begin graduate work at all or to complete a doctorate in science fields is very widespread, and was widely used in the past to justify denial of financial support to women graduate students (Bernard, 1964), thereby generating a self-fulfilling prophecy. One useful way of measuring suspected attrition is shown in Table 1.2, which compares the percent women among probable majors of entering college freshmen in 1968-69 with baccalaureate, master's, and doctoral degrees earned at appropriate subsequent intervals. Some field differences are apparent at once in the persistence rate through college: prospective majors in social sciences suffer a large decline by graduation, and a modest decline occurs in both mathematics and biological sciences, while the physical sciences and agriculture actually show a relative gain. At the master's degree level, significant attrition has taken place in mathematics and social sciences but several fields show slight gains, and in agriculture the likelihood of women's earning the degree is increased dramatically. With respect to the doctorate, by far the largest decrease occurs in mathematics, for an attrition rate of over 60 percent from the BA, most of which occurs after the master's degree. In chemistry the attrition rate is also large, amounting to about 25 percent from the baccalaureate. Biological and social sciences both show relatively small decreases between the baccalaureate and doctorate, while the probability of women's earning a Ph.D. stays essentially constant compared to men in computer sciences and physics but rises in engineering and agriculture. Such marked differences by field in persistence to the Ph.D. between men and women are not readily explained in any simple way. They do not correlate well, for example, with various measures used by Feldman (1974) to diagnose perceptions of sex equity for graduate students in various disciplines. This study provides evidence that in a large-scale national survey carried out in 1969, about 15-35 percent of both male and female graduate students in science fields believed that faculty do not take women seriously, a belief generally corro- borated by the corresponding faculty survey. However, this opinion was much less pronounced for mathematics than for chemistry, physics, most biological sciences, and several social science fields. The observed variations in persistence rates, both among fields and between the sexes, suggest that the nature of the choices made by men and women continues to be determined by somewhat different factors. In our pipeline model, there are several possible branches at each valve or degree level, such as immediate employment, a switch to a different aca- demic field, a switch to professional training such as law or medicine, and others. "Attrition" between the baccalaureate and the doctorate in certain fields (biosciences, some social sciences, chemistry) may, for example, partially reflect the very large increases in medical school attendance by women that followed the lifting of quota restrictions in these schools in the early 70s. 1.10
No clear explanation exists for the marked discrepancy in persis- tence rates between mathematics and other science fields; there is a resemblance to chemistry in that most of the drop occurs after the master's degree. Mathematics, however, is a small field compared to most other sciences. It is perhaps more important to focus attention on the high persistence rates in other science fields. 2. "Parity" at the doctoral level . Since a detailed discussion of women doctorates in science and engineering is presented in Chapter 2, this section will confine itself to assessing the degree of parity women have achieved at the Ph.D. level. It should be emphasized that the percentages of degrees earned by women in each field are the important determinant of their represen- tation in the labor pool, and therefore are the appropriate measure for career progress. Hence the actual proportions of degrees earned by women will serve as the reference standard throughout the remaining chapters of this report In this section, however, we apply two different measures to assess separately 1) the extent to which women's field distribution in a given Ph.Do cohort approximates that of men, and 2) the propensity of women baccalaureates in a given BA cohort relative to corresponding men to earn a Ph.D. after a time interval appropriate to the particular field. These measures, respectively termed PI1 (parity index) and PI2, are presented in Table 1.3, comparing results for 1970 and 1980 Ph.D.s. Table 1.3 shows that in 1970, when we consider the field distribu- tions of all doctorates, women were slightly more likely than men to have earned the degree in biosciences and much more so in psychology, but much less so in mathematics and least of all in physical sciences (PITS. With reference to the relevant baccalaureate pool at a field-specific appropriate earlier time (PI2), however, a somewhat different picture emerges, comparable to the persistence rates discussed aboveO Women baccalaureates in mathematics were only about one-quarter as likely as their male colleagues to proceed to a doctorate and in psychology that probability was 60 percent, with the other fields falling between these extremes. The reader should note that these ratios for mathematics are not highly reliable because they are based on numbers smaller than 100 (see Note c, Table 1.3~. By 1980 no change is observed in PI1 for physical sciences and the mathematics value has decreased, apparent evidence of a diverging sex distribution pattern in that field. The values for both biosciences and psychology are reduced compared to 1970. A dramatic change has occurred, however, in PI2 in all fields except mathematics; based on the relevant BA pools, there is a close approach to parity in bio- sciences and psychology, and the parity measure has nearly doubled in physical sciences to the .75 level, but has increased only slightly in mathematics. 1.11
TABLE 1.3 Parity trends for Ph.D.s-by broad field, 1970 and 1980 1970 . % Ph.D.s in field~to women PI PI2 1980 % Ph.D.s in field to women PI PI 1 2 rat Mathematics (7.8) (.59) (.27) (12.8) (~42) (.33) Physical sci. ~5~.~4 .40 .39 12.3 .41 .75 Biological sci. 14 3 1.08 .51 28.0 .92 .94 Psychologyd 22.3 1.68 .60 42.3 1.40 .97 aPI - Pexcent:wo~en Ph.D.s in field ~ . ~ Percent wome--n Ph.D.s in all fields bp _ Percent women- Ph.D.s in field In - _ Percent women BAs in field y years earlier where y = 6 years for physical sciences = 7 years for biological sciences = 8 years for mathematics and psychology Bracketed figures indicate percentages based on absolute numbers of Ph.D. below 100, where indices become very unstable due to fluctuations. d s Because of earty d~ffe\rences in data collection, aggregate figures for the broad field of social sciences are not available. SOURCE: Computed from National Center for Education Statistics, 1979 and 1980; national Academy of Sciences, 1981. 1 12
These patterns are clear evidence of strong convergence of men's and women's educational choices and decisions in all science fields except mathematics. It should be noted that a similarly converging pattern is developing in engineering and computer sciences, although the numbers of women Ph.D.s are too small to warrant separate presenta- tion here. These distribution ratios demonstrate among other things the importance of taking account of general education history when we attempt to explain women's low participation in science; in particular, the size of the appropriate baccalaureate pools has a marked influence on determining the degree of parity to be expected. It seems evident that until women achieve general parity at the baccalaureate level (i.e., 52 percent of all BAs), an event expected within the next five years or so, representation parity at the Ph.D. level cannot occur. Parity at the BA level (which was almost achieved just before World War II, although the data have not been presented here) depends primarily on two factors: the absence of formal sex distinctions in admissions and equality of access to financial support. Unless these conditions can be met, it is futile to expect full participation of women in science, and hence to have available an expanded pool of sci- entists to fill national needs. While formal admissions quotas for women have been eliminated in most institutions, the situation with respect to equal financial support may be more problematic. In the years immediately after World War II, the proportion of male college students rose from 55 percent to over 70 percent, and about 70 percent of all the men were supported by the G.I. Bill (Olson, 1974~. AS a result of this subsidy, higher education became accessible to men with the requisite ability regardless of their economic status, while women had to continue to finance their own education, thus in effect restrict- ing the numbers who could attend college. This situation was exacer- bated by the common practice of awarding women less financial aid from institutional sources as well; as late as 1969-70, institutionally- administered grants to male college students averaged $671 and those to women only $515 (Haven and Horch, 1972), a difference of 32 percent. Extensions of the G.I. Bill for Korean and Vietnam war veterans had ef- fects similar to the original bill but much reduced in total numbers and value. If the efforts to increase military recruiting through the promise of future educational benefits that are now under way in the Congress prove successful, such a policy will again adversely affect women's rela- tive access to higher education and hence to science careers. Even if future enlistments in the military forces remain open to women, it is un- likely that the present low quotas and higher qualifications required of women will be changed: on the other hand, there is reason to believe that scientifically talented young people of either sex would not enlist in sig- nificant numbers except on the case of a national emergency. Nonetheless, educational benefits as a reward for military service would again produce an artificial long-term dominance of men in higher education, such as occurred in the period following World War II, and hence revived the damaging perception that women are somehow unfit for advanced work. 1.13
Therefore this committee recommends that the Congress explore measures . . . to ameliorate the inevitable sex bias in higher education that will again result from a renewed G.I. Bill. The purpose of raising this question is to emphasize our concern not simply with the numerical dis- parities that result from a sex-biased support pattern but particularly with the consequent limitations on access for talented women from less than affluent families. The increases in women's propensity to prepare for careers in science and engineering over the last decade are too rapid to have been caused by some fundamental change in either female aptitudes or psychological makeup. Such a change is likely to require a far longer time period. It is much more probable that the increases have occurred as the result of two related factors: the ending of formal sex discrimination in institutions of higher education, in large part by the various civil rights statutes, and the consequent perception by women that since these statutes guarantee a measure of equality it is now relatively far more rewarding to make the personal and financial investment that is required to prepare for careers in science and technology. The PI2 ratios allow us to make some modest predictions of the proportions of women to be expected in Ph.D. pools in the future. If we assume that the rates at which women perceive science fields to offer an equitably rewarding future do not change from the 1980 levels, we would expect about 18 percent women doctorates in physical sciences in 1986 and about 40 percent in biosciences in 1987, but only 14 percent in mathematics by 1988. If women's propensity to complete doctorates in physical sciences continues to rise, on the other hand, at the same rate as in the recent past, women in these fields would attain statistical parity in the late 1980s at about 25 percent of Ph.D.s. Numerical parity is likely to be attained only after it has occurred at the baccalaureate level, which would happen in about the mid-199Os at the current rate of growth. The fact that science participation by high school girls is currently increasing gives reason to believe, however, that the growth of the female BA pool will accelerate in the near future, other factors remaining equal. The situation in two fields of great current interest, engineering and computer sciences, is still much more uncertain. In engineering women started from a near-zero base a decade ago and were 10 percent of baccalaureates and 3.6 percent of Ph.D.s in 1980. The high growth rates for women undergraduate engineers became most marked in the last few years and will not be reflected among Ph.D.s until the late-1980s. In computer sciences (where separate records are not avail- able before 1971) women BAs were 13.6 percent in 1971 and 30.2 percent in 1980, but the proportion of Ph.D.s has remained essentially constant at about 9.5 percent. It is probable that in this field also the rate of growth will accelerate during the 1980s, based on baccalaureate figures. 1.14
Two kinds of conclusions emerge from these considerations. The first is that although the participation rates of women in mathematics graduate degrees remain something of a puzzle, in most science fields the flow through the pipeline is accelerating at a rate that leaves little doubt of either the fitness or inclination of women to under- take professional careers in these disciplines. We have noted elsewhere (see Chapter 2 and also Climbing the Academic Ladder, pp. 23-32) that the quality of women Ph.D.s is at least equal to that of men in all objective respects. If the pool of women scientists were to be expanded to the size of the male pool, it is likely that the quality of the two would be identical. In principle, therefore, the pool of professional scientists can be expanded very considerably at no loss of quality; more important, the quality of that pool of constant size can also be raised by giving more encouragement to women than they have received so far. In- termediate options also exist, of course. It should be noted here that while various organizations that work with women graduate students (for example, Higher Education Resource Services) report such students do not currently perceive much active sex discrimination, in contrast to a decade ago, they also do not perceive the same encouragement as that given to male students. The second conclusion is that the data presented here lend new emphasis to the necessity for adequate science and math education of girls at the pre-college level. The necessity for equal science and mathematics education of girls and boys throughout the school years cannot be overemphasized, and school science programs must insure that girls participate equally with boys. There are a number of programs aimed at female high school students, and their teachers and counselors, conducted on a local or regional basis, that have had positive results in encouraging interest in mathematics and providing career information. As an example, we note the Equals program in northern California which has worked with secondary school teachers to promote strategies for encouraging girls to continue with mathematics. The general topic of the quality of pre-college education in these fields has already aroused much concern on the part of the National Academy of Sciences and the National Science Foundation. We wish to record our urgent recommenda- tion that potential new programs growing out of this concern be developed with careful regard to the equitable participation of female and minority . . . . . students at all levels and from the beginning. In particular, Congressional authorization of such pro~rams_and potential state initiatives should include explicit support for such participation in the distribution of public funds. It is important to insure that every young person has the opportunity to assess his or her interest and talent for advanced study in mathematics and the sciences and for the pursuit of professions based on these fields, in order to insure that in the long term the best talent will be available in the advancement of science and in its application to national needs. 1.15
