Janice Hudgings, Assistant Professor, Physics Dept., Mount Holyoke College

Sheila Humphreys, EECS Department, University of California at Berkeley

Patrick Hernan, EECS Department, University of California at Berkeley

July 1999

The number of women graduate students enrolled in the Electrical Engineering and Computer Science (EECS) Department at UC Berkeley dropped by 33% between 1993 and 1998. This trend is paralleled by a substantial reduction in the fraction of incoming women enrolled for the Ph.D. as opposed to the Master’s degree. In order to investigate the origins of this trend, we examine admission, enrollment, and retention of women graduate students admitted during the six years from 1993-1998. From this data, we conclude that the sharp reduction in the number of women graduate students arises from the combination of a 50% drop in the yield of admitted women students who chose to enroll in the EECS Department from 1993-1998 and a precipitous decline in degree completion rate. Of the 40 women who enrolled in the EECS doctoral program between 1993 and 1995, 48% of these women left the department without completing the Ph.D. If the department is committed to maintaining a diverse pool of graduate students, the declining yield must be addressed by aggressive recruitment of admitted women students. Furthermore, we urge the EECS Department to adopt an exit questionnaire in order to track degree completion and to examine reasons why students leave the EECS graduate program. Routine analysis of enrollment patterns and retention of women in graduate degree programs by the EECS Department will facilitate an informed discussion and appropriate intervention.

The underrepresentation of women in engineering has been well-documented, with current studies estimating the participation of women in engineering at less than 12% nationwide.¹ This trend is consistently reflected at all levels of education: bachelors, masters, and doctoral. Women earned only 10% of engineering doctorates in 1995, whereas they attained over 40% of Ph.D.s in the biological sciences.² Within UC Berkeley, 17% of enrolled undergraduates in the Electrical Engineering and Computer Sciences (EECS) Department are women. At the graduate level the percentage of women EECS students has been steadily declining in recent years, from a high of 17% in 1993 to 15% in 1998, reflecting both a 50% decline in the rate of enrollment of women graduate students and also a sharp decline in the retention of women doctoral students.

Notably, low participation at the graduate level by women has been cited as a primary obstacle to gender equity at advanced levels within engineering.³ Masters degrees are now widely expected professional qualifications for EECS students, and completion of a doctorate is prerequisite to obtaining an academic faculty position.

Our motivation for this study is to identify emerging trends in admissions, enrollment, and degree completion by women graduate students in the Electrical Engineering and Computer Sciences Department at UC Berkeley. We examine admission, enrollment, and retention of women graduate students admitted during the six year period 1993-1998. This period spans the two years immediately prior to and the two years immediately following the passage of both Special Provision 1 by the Regents of the University of California in July 1995 and also Proposition 209 by the voters of California in 1996, banning race and gender preferences in admissions and retention programs. We use these enrollment and retention data to assess existing strategies, such as the department parent policy and the Computer Science Reentry Program, for increasing the number of women in the EECS graduate program. Based on this information, we make recommendations to the EECS Department for recruitment, admission, and retention of graduate women.

The Department of Electrical Engineering and Computer Sciences has made a concerted effort since the mid-1980’s to increase diversity in the graduate population by actively addressing both the recruitment and retention of women graduate students. In 1983, the EECS Department instituted the Computer Science Reentry Program in order to address concerns about the shortage of potential women applicants equipped with the computer science background necessary to apply for admission into the computer science graduate program. This program has enabled women with nontraditional backgrounds to complete the requisite undergraduate coursework to obtain admission on a fully competitive basis to graduate computer science programs. In 1985, the Dean of the College of Engineering (Karl S. Pister) instituted a graduate recruitment program directed toward women and underrepresented minority students. The EECS Department quickly followed suit by creating a staff position and initiating active fundraising and recruitment. This program consisted of outreach travel by faculty and staff, special publications, a letter and telephone campaign between current women students to admitted women, and a program of coordinated, subsidized campus visits. Targeted fellowships designated for women were established, and oversight carried out to ensure that all admitted women students received financial aid in the form of a fellowship or a teaching or research assistant appointment.

The Berkeley Graduate Division data tracking successful completion of the Ph.D. show a disparity in completion rates. Of the students who enrolled in the EECS Department from 1984-1986, only 43% of women doctoral students obtained a Ph.D., whereas 64% of their male colleagues succeeded in achieving this goal. In addition, women completing a Ph.D. in Computer Science from 1993-98 took, on the average, about one semester longer to complete the CS doctorate (7 years) as compared to male students (6.6 years). The mean time to complete the doctoral degree in EE, however, was identical for men and women (6.1 years for women as compared to 6.2 years for men) in the same period.

A detailed survey of Ph.D. attrition among EECS students who entered the program between 1981-1991 and left without completing an originally intended Ph.D. was completed in 1996 by the graduate women’s group WICSE (Women in Computer Science and Electrical Engineering). They found that the most common reason cited by both men and women for leaving the graduate program was "changed career goals and personal reasons,’’ followed closely by lack of academic mentoring and lengthening times to degree.⁴ This problem has been addressed with a two-fold approach. First, the EECS Department continues to maintain strong administrative and financial support for WICSE, a student-run networking and support group for women graduate students in EECS. Second, in 1996, the department instituted a remarkably proactive student parent policy in order to address concerns expressed by women graduate students that the long time to degree completion overlaps with primary child-bearing years.⁵

With these programs in place, the percentage of women in the graduate program increased steadily from 12% of the total graduate enrollment in 1986 to a peak of 17% in 1993. Similarly, the number of doctoral degrees earned by women in electrical engineering and computer science increased in that decade. Recently, however, both enrollment and completion rates have declined precipitously; in particular, we note a significant reduction in the completion rate of doctoral degrees by women in the EECS Department.

Data on the applications, admissions and enrollment of women in the graduate program in EECS were collected for 1993-98. The number of applications from women during this period increased by 35%, from 224 in 1993 to 303 in 1998. Applications from women reflected the fluctuations of the overall applicant pool; overall applications increased 18% in the same period. Furthermore, this trend continued after the passage of Proposition 209; between 1997 and 1998 women’s applications increased by 32%, while overall applications increased by 21%. Thus, while only 15% of the applications in 1998 to the EECS graduate program were from women, the absolute number of women applicants is increasing steadily. This indicates that the size of the applicant pool is not causing the declining enrollment of women graduate students in EECS.

Likewise, the rate of admissions for women during the six years from 1993-1998 is relatively high. In each year except 1996, the admissions rate for women exceeded the overall admissions rate, as shown in Table 1 below. Both prior to and after the passage of Proposition 209, women were admitted at slightly higher rates than were men. For example, the overall rate of admissions in 1998 was 12%, whereas the rate for women was 15%. Although outside the bounds of our study, it is notable that the rate of admissions for women for Fall 1999 dropped to 10%, as compared with 13% overall.

Alarmingly, despite strong rates of applications and admission of women, the "yield" of women graduate students, defined as the number of women who chose to accept the EECS Department’s offer of admission, has fallen dramatically between 1993 and 1998. In 1997 and 1998, the two years in which Proposition 209’s effect began to be felt, the percentage of women accepting Berkeley’s offer of admission to EECS dropped far below the overall yield rate for the department, as shown in Figure 1. In 1993, nearly half of all women who were admitted accepted the EECS Department’s offer; by 1998, the percentage dropped to a low of 24%. In that same year, the overall pool of admitted applicants accepted at a rate of 39%. The shrinking "yield" obviously has contributed to both a lower number of new graduate women as well as the total number of women graduate students in EECS. In 1993, 97 women were enrolled in EECS; in Spring Semester 1999, 65 women were registered. This decline in yield may reflect negative attitudes toward Berkeley or increasing industry competition for qualified students at the bachelor’s degree level. Because of the confusion in legal interpretation of Proposition 209, it is also possible that Berkeley faculty, staff and students recruited less aggressively. This declining yield is dramatically reflected in the number of incoming women enrolling as Ph.D. students. The percentage of incoming women enrolled for the Ph.D. dipped from a high of 93% in 1994 to a low of 31% in 1997. In each of the years from 1993 to 1996, between 11 and 14 women enrolled as Ph.D. students. This number dropped by more than 50% after the passage of Proposition 209, with 5 new women Ph.D. students enrolling in EECS in 1997 and 7 in 1998.

Figure 1. Comparison of EECS Graduate Program Yield Rates

In addition to tracking admissions and enrollment data, we investigated the retention rate, or successful degree completion, of women in the EECS graduate program. A total of 88 women entered the EE and CS graduate programs from 1993 to 1998. We tracked individual students in this cohort by their year of entry into the program and their degree outcome, defined by whether each student completed the degree for which she was originally admitted.

The average time to complete a Ph.D. is 6 years for Electrical Engineering and 6.8 for Computer Science. Predictably, of the of students who entered in 1993, only three had completed a Ph.D. by 1998. One student from the 1994 cohort completed the Ph.D. by Spring 1999. To date, 48 of the 87 women in the cohort have successfully completed Master’s degrees. The numbers of both Ph.D.s and Master’s degrees awarded to women in the cohort will continue to increase as the later classes progress.

The retention rate of women in the EECS graduate program, as defined by completion of the originally intended degree, is illustrated by Figure 2 below, showing degree outcomes for women entering the program in 1993-1995. The data for the women entering in 1996-1998 are omitted, as these students have not had time to complete a substantial portion of their graduate work. Of the 40 women admitted as Ph.D. students during these three years, 18 left the program with a Master’s degree and one left with no degree. Furthermore, the rate at which women Ph.D. students in EECS left with a Master’s degree climbed steadily, doubling between 1993 and 1995. Three of these students were National Science Foundation Graduate Fellows, who were selected for showing superior academic promise and were entitled to three years of financial support for the Ph.D. Comparable data, not currently available, need to be collected and analyzed for a parallel cohort of men in order to determine whether male doctoral students left the Ph.D. program at the same rate.

It is important to note that there are a number of confounding factors influencing why these women left the EECS doctoral program. Only three of the women in our cohort left for academic reasons such as not passing the preliminary exam or not being encouraged to take the preliminary examination because of a low GPA. None of the women transferred immediately to another graduate school. The vast majority of the women doctoral candidates in our study who did not complete the Ph.D. left the university for opportunities in industry, perhaps reflecting the unprecedented recent demand and very competitive job market in Information Technology. EECS graduates with a Masters degree now command average starting salaries of $65,000, and recruitment pressure from industry is intense. Students report anecdotally that the current job market for EECS graduates decreases the motivation to pursue a doctoral degree. The lack of exit interview data from the EECS Department precludes determining whether students left the Ph.D. program because of the strong job market or because of the "personal reasons" or the academic hoops and "lack of academic mentoring" described in the earlier 1996 study.⁴

Figure 2. Degree Outcomes for Women Graduate Students in EECS

The EECS Department has made a concerted effort to provide strong academic support for its graduate students, with a variety of programs implemented by the Graduate Matters Office. The major academic challenges and hurdles affecting graduate students: new student orientation, advising, annual review, and the Preliminary and Qualifying Exams, have been evaluated and refined since the mid-1980’s in response to student input. Annual student reviews are now structured as a dialogue, giving students the chance to offer feedback to their faculty advisor. Graduate students meet with their advisors for what has evolved into a two-way review at least once a year and then receive a review letter. The coordination of financial support for graduate students is handled by the faculty Vice Chair for Graduate Matters, and all students now enjoy financial support in the form of a research assistantship, teaching assistantship, or fellowship. Student organizations have established a tradition of weekly social hours, paid for by the EECS Department to promote a sense of community. The voice of graduate students is heard at occasional "Town Meetings" at which any student may raise issues in the presence of the Vice Chair for Graduate Matters, and at the EECS Faculty Retreat, at which graduate student groups make an invited presentation.

These extensive support programs are reflected in the results of the Graduate Division’s exit questionnaire, which in 1991 showed that 89% of the exiting EECS Ph.D. students were either "satisfied" or "very satisfied" with the "overall quality of graduate level teaching." Similar data from graduating Ph.D.’s were again compiled in 1997; doctoral students graduating during the past 5 years ranked EECS as tenth out of 80 Berkeley departments on overall satisfaction on a composite of five key questions: "quality of graduate teaching; level of financial support; departmental advising/guidance, professional relationship with one’s doctoral supervisor, and faculty efforts and assistance in finding professional employment." The students ranked the Department highest on "level of financial support" and lowest on "relationship with Ph.D. supervisor" and "finding employment." These findings echo student perceptions of concerns about lack of guidance from the research advisors reported in the WICSE survey cited above.⁴

In addition to general retention efforts, the EECS Department has made a concerted effort to improve the rate of retention of women graduate students. Previous research has shown the importance of role models and achieving a "critical mass" of students for the retention of women in engineering. EECS women have barely achieved critical mass, defined as 15%. In addition, studies show that graduate students in mentoring relationships demonstrate the best progress toward degree completion.⁶ Therefore, a large focus of the retention program for women centers on support for WICSE, the graduate women’s group, to promote peer contact.⁷ Entering women are paired with more senior student "Big Sisters." The EECS Dept. funds the students’ weekly lunches, and assists WICSE to make contact with women faculty speakers. Those earning graduate degrees are honored at the end of each year at a special event. Funding is available to enable EECS graduate women to attend relevant conferences such as the Grace Murray Hopper Celebration of Women in Computing, or the Computer Research Association’s Conference for Women in Research. Deliberate effort has been expended to maintain a database of women doctoral graduates on a web page (http://hera.eecs.berkeley.eecs/Programs/gradwomenDDBS.html) to increase contact between current women and alumnae, and women graduate students are paired in formal mentoring relationships with alumnae in industry and academia.

In addition, increasing the number of women faculty has been an unarticulated aspect of retention for women in the EECS Department. Professor Susan Graham was the sole female faculty member in the EECS Department from 1976-1988, when Professor Avideh Zakhor was appointed as the first electrical engineering faculty member. To fill in the gap, the College of Engineering and the EECS Department invited a series of distinguished women academics and industry researchers between 1985-1995: Professor Mildred Dresselhaus, Institute Professor at MIT; Fran Berman, IBM Research, Ruth Davis, National Academy of Engineering, and Dr. Betsy Ancker-Johnson, General Motors. In addition, women were encouraged to come to EECS as part of the NSF "Visiting Women Professor Program" (now discontinued): Dr. Jeanne Ferrante, IBM Watson Research; Professor Mary Lou Soffa, Univ. of Pittsburgh; and Dr. Ljiljana Trajkovic, Bellcore. As of 1999, five women hold faculty appointments in EECS: two in Computer Science and three in Electrical Engineering. All are tenured. Nevetheless, no women were interviewed for faculty positions in Computer Science from 1997-99, and only one woman interviewed for Electrical Engineering. The prospects of more women joining the faculty in the near future look weak.

With the establishment of an EECS Undergraduate Matters Office in 1997, a new focus on recruiting freshmen women was developed, including personal outreach to admitted students by faculty via the Internet as well as telephone and letter campaigns. In 1998, the highest number (32) of freshmen women ever entered the Department. Undergraduate women now represent 17 % of the total enrollment and eventually, their growing numbers may contribute to the graduate population. A major program is underway to increase the number of undergraduate women who participate in research, both for retention and to motivate them for graduate study.

In addition to these more traditional efforts at increasing the pool of potential graduate students, the Computer Science Division of the EECS Department initiated the Computer Science (CS) Reentry Program in 1983. The CS Reentry Program’s initial goal was to increase the number of women in the doctoral pipeline in computer science, by offering admitted women from "nontraditional" educational backgrounds participants a unique chance to replicate the knowledge acquired in an undergraduate CS major. These reentry students then apply to graduate programs in Computer Science on a fully competitive basis. Since UC Berkeley does not offer part-time study, the Reentry Program is unique that reentry students have been offered the chance to take courses at a slower pace than the standard timetable imposed on undergraduates. After ten years, the Computer Science faculty reviewed the program in 1994 and voted to expand the program’s goal to include increasing the number of women who earn master’s degrees in Computer Science.

To date, 169 students have participated in the Reentry Program, of whom 162 are women and 8 are underrepresented minorities. The vast majority of Reentry students have either gone immediately into the technical workforce or have gone on to earn a Master’s Degree in Computer Science. In addition, during the 15 years of operation of the program, four graduates of the CS Reentry Program have earned Ph.D.s in Computer Science, three of which were awarded at Berkeley. Since the Reentry Program’s inception, 22 students have matriculated in the EECS graduate program, with four reentry women entering the Berkeley graduate program from 1993-98. Three of those students are now progressing toward the Ph.D. A more meaningful measure of the program’s success is the 38 reentry students who have earned M.S. degrees in Computer Science at Berkeley or elsewhere; three of these students were underrepresented minorities. Eighteen CS Reentry students completed M.S. degrees at Berkeley; 4 at MIT, and 7 at Stanford. Clearly, the CS Reentry Program has contributed substantially to the graduate pipeline at UC Berkeley and elsewhere.

As a consequence of Proposition 209, the Computer Science Reentry Program has been forced to suspend the program in its present form, because it is considered an "educational benefit" based on gender or ethnicity.

In response to concerns expressed by women in EECS that graduate studies conflict with primary child-bearing years, creating additional financial and academic pressures, the EECS faculty approved an innovative Parent Policy in 1995. This policy provides extra time to graduate student parents for meeting degree requirements such as the Preliminary Examination and the Qualifying Examination, due to their parental obligations. Student parents are permitted an extra year for both of these milestone exams. Expectations for research productivity immediately before and after the birth of a child are reduced without the loss of funding from a Graduate Research Assistantship. In addition, the policy allows for supplemental financial aid based on documented need.

Funding for this new parent policy was solicited from industry, but due to a weak response and lack of additional extramural funds, the supplemental financial aid aspect of the policy has not been implemented. Although officially sanctioned, no money is available to provide financial supplements to needy student parents. Current and prospective students are not sufficiently aware of this policy. The effect of the policy is still unknown. An important campus-wide effect of the innovative departmental policy, however, was its influence on the Berkeley campus administration, which appointed a task force to consider the needs of student parents and adopted a similar policy in May 1998 for the entire student body.

In conclusion, the number of women enrolled in the EECS graduate program has declined by 33% from 1993-1998. This is paralleled by a substantial decline in the fraction of incoming women enrolled for the Ph.D. as opposed to the Master’s degree. Clearly, if the EECS Department is committed to maintaining a diverse pool of graduate students, this trend needs both to be monitored on a continuing basis and to be actively addressed.

Several potential reasons for this decline in the number of women graduate students have been investigated. While only 15% of the applicants to the EECS graduate program in 1998 were women, the absolute number of applications from women increased by 35% from 1993 to 1998. Furthermore, the rate at which women applicants were admitted to the EECS program exceeded the overall admission rate during this period, with the exception of the 1996 admissions. For this reason, we conclude that neither the number of applications from women nor the rate of admissions is the cause of declining enrollment of women graduate students in EECS. However, it is important to emphasize that while applications and admission are not responsible for the recent decline in the numbers of women graduate students, both of these factors are essential to any future effort to increase the number of women in the graduate program. The EECS Department cannot hope to achieve parity of opportunity for women graduate students if the fraction of applications from women remains at 15%, and the 1999 drop in the rate of admissions of women (10%) to below the overall admission rate (13%) is alarming.

The 33% reduction in the number of women graduate students in EECS from 1993-1998 can be attributed to a 50% reduction in the yield of admitted women during these years, in conjunction with a significant decrease in the degree completion rate by women. While it is unclear whether the declining yield results from increasing competition from industry or negative attitudes towards UC Berkeley in particular, it is essential that the EECS Department aggressively recruit admitted women graduate students if the enrollment trend is to be reversed.

Perhaps most remarkably, of the 40 women who enrolled in the EECS doctoral program between 1993 and 1995, 48% of these women left the department without completing the Ph.D. All but one of the women who left without completing the Ph.D. left the department with a terminal M.S. Furthermore, this is an increasingly common phenomenon; the number of women doctoral students who left with a Master’s degree doubled between 1993 and 1995. It is unclear whether this trend is specific to women students, as no similar study has been completed of degree outcomes of male graduate students. However, regardless of whether this reduction in completion rates of doctoral degrees is gender specific, it is certainly cause for concern. Furthermore, the reasons why women doctoral students are increasingly leaving without completing the Ph.D. are unknown; it is uncertain whether this phenomenon is due to a remarkably vibrant job market or to the "personal reasons" and "lack of academic mentoring" cited in the 1996 study.⁴ We recommend that the EECS Department institute a policy of exit questionnaires in order to track degree completion and to investigate why students leave the graduate program.

Programmatic initiatives within the EECS Department to improve both the rate of enrollment and the rate of retention of women graduate students have been of varying degrees of success and have been hampered by Proposition 209. In particular, the department’s strong fiscal and administrative support for WICSE, the graduate women’s group, ensures a continuing source of peer role models and mentoring initiatives for women graduate students, which has been shown to improve degree outcome.⁶ However, efforts to increase the number of women faculty have become somewhat halting in recent years, resulting in a shortage of more senior women role models. The Computer Science Reentry Program has successfully increased the pool of qualified women applicants to the graduate program at both UC Berkeley and elsewhere; from 1993-1998, 12% of the women enrolled in the computer science graduate program at UC Berkeley were participants in the Reentry Program. The elimination of this program due to Proposition 209 will have an immediate effect on the number of available women applicants. We suggest investigating alternative, private funding to support its continuation outside the constraints of Proposition 209. Policies sympathetic to student parents, such as the Parent Policy, provide encouragement but have not been fully leveraged for recruitment or retention. Easing the way for students who wish to rear a family at the same time they are pursuing a graduate degree will enable more women to pursue graduate degrees, and we recommend active implementation of this policy.

In order to reverse the alarming 33% decline in the number of women enrolled in the EECS graduate program, the department must address both the declining yield of admitted women students and the sharp reduction in degree completion by women doctoral students. We urge the Department to undertake a more ambitious study, in which interviews are conducted to probe the reasons why students abandon degree goals. We recommend that the EECS Department routinely track cohorts of entering students and the degrees they earn in relation to their originally expressed intentions. Routine analysis of enrollment patterns and, in particular, retention of women in graduate degree programs by the EECS Department will facilitate an informed discussion and appropriate intervention.

We wish to thank the Graduate Division, University of California at Berkeley, for partial support of this study. We acknowledge with appreciation the assistance of Judy Sui, Graduate Division, Mary Byrnes and Ruth Gjerde, EECS Graduate Matters, and Barbara Hightower, Letters and Science Advising and Reentry Coordinator, and thank Professor Andrew Neureuther for his helpful comments.

1. "Science and Engineering Indicators," National Science Board, 1998. http://www.nsf.gov/sbe/srs/seind98/start.htm
2. K. Olson, "Despite Increases, Women and Minorities Still Underrepresented in Undergraduate and Graduate S&E Education," National Science Foundation Data Brief, NSF 99-320, January 15, 1998.
3. Women, Minorities, and Persons with Disabilities in Science and Engineering: 1996, National Science Foundation, September, 1996, p. 54.
4. L. Kamas, C. Paxson, A. Wang, and R. Blau, "Ph.D. Student Attrition in the EECS Department at the University of California, Berkeley," unpublished paper, UC Berkeley, 1996.
5. EECS Graduate Student Notes 1998-99, University of California, Berkeley, pp.83-4.
6. G. E. Girves and V. Wemmerus, "Developing models of graduate student progress." Journal of Higher Education, 59 (2), 163189, 1988.
7. S. Humphreys, "The Role of Women Graduate Students," Conference Proceedings from Bridging the Gender Gap, Carnegie Mellon University, 1996, pp. 33-40.