Back to May 2002 CRN Table of Contents

[Published originally in the May 2002 edition of Computing Research News, pp. 2, 11.]

Expanding the Pipeline

By Barbara M. Moskal

This is a shorter version of an article that will appear in the June 2002 issue of the ACM SIGCSE Bulletin, inroads, which focuses on women in computing. The issue is edited by Tracy Camp, Colorado School of Mines.

Introduction

As new doctorates complete their degrees in U.S. academic institutions, they are asked to respond to the Survey of Earned Doctorates (SED), an instrument that is designed to collect self-report information concerning graduate school experiences and future employment. The SED seeks to be a census of all doctoral recipients in the United States. The survey is sponsored by the National Science Foundation, the National Institutes of Health, the U.S. Department of Education, the National Endowment for Humanities, the U.S. Department of Agriculture, and the National Aeronautics and Space Administration. Copies of the specific questions that appear on the SED are available in the appendices of the yearly final reports (Sanderson, Dugoni, Hoffer, & Myers, 2000; Sanderson, Dugoni, Hoffer & Selfa, 1999; Sanderson & Dugoni, 1998). Detailed reports and statistical tables concerning science and engineering can also be found in a separate series of reports (Hill, 2001a; 2000b; 2000; 1999; 1997; NSF, 1996; 1995).

This paper summarizes the results of an analysis of a subset of data that was collected through the SED. The focus of this paper is on women who completed doctorate degrees in computer science between the academic years 1990-91 and 1999-2000. A more extensive version of this report that includes both the methodology and statistical summaries is available at Moskal (2002). Although computer science doctorates are defined here to include doctorates in information science (WebCASPAR, 2001), prior research (Camp, 1997) suggests that the percentage of women completing doctorates in information science is low compared with computer science. This article is divided into the following subsections: National Trends in Doctorate Degrees, Experiences Completing the Doctorate, Experiences After the Doctorate, and Conclusions.

National Trends in Doctorate Degrees

Between the academic years 1990-91 and 1999-2000, the percentage of degrees awarded to women in computer science, science and engineering, and all degrees has increased from 14.63% to 16.49%, 28.86% to 36.17%, and 36.96% to 43.80%, respectively. This results in a ten-year rate of increase of 12.77% for computer science, 25.34% for science and engineering, and 18.51% for all degrees. If these rates of increase continue on a per-year basis, women will reach parity with men with respect to earned doctorates in science and engineering by 2012-13 and across all fields by 2007-08. In computer science, parity with men will not be reached until the academic year 2087-88-more than 80 years from now.

Within science and engineering and across all fields, the increase throughout the ten years of interest has been steady. The trend in the percentage of computer science degrees that were awarded to women is far less consistent, displaying both increases and decreases during this period. Given the consistent increase in science and engineering degrees and all degrees awarded to women, it is reasonable to assume a continued, consistent increase in the next several years in these areas. Since the awarding of computer science degrees has displayed no clear trend, the assumption made here of a steady increase in the years to come is optimistic. Based on a trend that began in 1997-98, a decrease in the near future is more likely.

The next statistic concerns women who are U.S. citizens or permanent residents. Over the same ten-year period, there was a consistent increase in the percentage of science and engineering (35.17% to 41.69%) and all doctorate (43.08% to 48.89%) degrees that were awarded to women. The rate of increase for this period across science and engineering and across all fields is approximately 18.54% and 13.48%, respectively. The trend in the percentage of female U.S. citizens or permanent residents to receive doctorate degrees in computer science has been far less consistent, and displayed a 9.91% decrease across the ten-year period (20.84% to 18.78%). If these trends continue, women who are U.S. citizens or permanent residents and who receive doctorates in science and engineering and across all fields will reach parity with men in 2009-10 and 2001-02, respectively. For women who are U.S. citizens or permanent residents, parity with men in the attainment of doctorate degrees in computer science is not anticipated.

Experiences Completing the Doctorate

Overall, women spent more time enrolled in graduate school than did men over the ten-year period of interest. The mean difference between these values ranged from 0.4 to 1.4 years. The median time enrolled for women was greater than that of men for all but one academic year, 1995-96. During the 1995-96 academic year, men and women both had a median of 7.2 years invested in their graduate education. The median additional time required for women across the ten-year span ranged from 0 to 1.3 years.

While completing a graduate education, students require financial support for both living and educational expenses. One question on the SED asks respondents to indicate their primary source of financial support during their graduate education. Women were more likely than men to rely on personal funds (e.g., savings, personal loans and family support) (female, male difference: 7.03%) as their primary source of support. Women were also more likely to report fellowships, scholarships, or dissertation grants as their primary source of support than were men (female, male difference: 2.38%). Men, on the other hand, were more likely than were women to rely on research assistantships, trainingships, or internships as their primary source of support (male, female difference: 9.65%).

Similar results emerged when the data were restricted to U.S. citizens or permanent residents. Across the ten years, women were more likely than were men to rely upon personal funds as their primary source of support (female, male difference: 6.16%). Women were also more likely to rely on fellowships, scholarships, or dissertation grants as their primary source of support, compared with men. This difference, however, was slight (female, male difference: 1.47%). Men continued to report research assistantships, trainingships, and internships as a primary source of funding at a higher rate than did women (male, female difference: 10.83%). One notable difference from the previous results is that women who are U.S. citizens or permanent residents were slightly more likely to rely on teaching assistantships for their primary source of support than were their male peers (female, male difference of 1.91%).

Experiences After Completing the Doctorate

The next set of statistics addresses what male and female computer science doctorates did after completing their degrees. Individuals with unknown plans accounted for less than 10% of respondents and were eliminated from further analysis. Across the ten-year period, the majority of women (88.97%) and men (86.18%) had definite plans within the United States. Individuals who indicated that they had definite plans within the United States were asked to indicate whether these plans were to continue study or to become employed. The majority of these individuals indicated that they would become employed.

The majority of computer science doctorates who had definite employment plans in the United States either entered an academic position at a college or university or entered industry upon graduation. The other possible occupations accounted for less than 10% of those who had definite employment plans in the United States. Across the ten-year period, 53.92% of the women, compared with 35.00% of the men, entered academic positions. When this analysis is restricted to U.S. citizens or permanent residents, 55.85% of women and 37.89% of men entered academic positions. Men were more likely than were women to enter industry positions upon graduation. Across the ten-year period, men reported entering industry positions at a rate of 56.71%; women reported entering industry positions at a rate of 38.24%. When this analysis is restricted to U.S. citizens or permanent residents, men entered industry positions across the ten-year period at a rate of 51.04% and women at a rate of 34.95%.

Conclusions

This section briefly summarizes the major findings that were presented above. Each summary is followed by a question that is left for future research.

• The percentage of women who acquire doctorate degrees is gradually approaching parity with the percentage of men who acquire doctorate degrees. This is true regardless of whether the data included all doctorate degrees earned or are restricted to science and engineering doctorates. Yet, the percentages with respect to computer science doctorates are very different. At the current ten-year rate of growth, parity across men and women with respect to computer science doctorates will not occur for more than eighty years. Since the percentage of female U.S. citizens or permanent residents who are earning doctorates in computer science is decreasing, parity for men and women in this subgroup is not anticipated. Why, despite the growth in female doctorates in science and engineering, are women not pursuing doctorates in computer science?
  
• Women reported a higher mean and median amount of time enrolled between the completion of their bachelors' degrees and the completion of their doctorate degrees. Closer inspection of this data suggests that women require up to a year longer than men to complete their graduate education. This finding was consistent across the ten years of interest, and was true regardless of whether the data included all doctoral recipients or whether the data were restricted to U.S. citizens or permanent residents. Does the additional time required of women in the pursuit of a computer science doctorate discourage female participation in computer science?
  
• Although women appeared to have the advantage over men with respect to two sources of primary support -fellowships, scholarships and dissertation grants and teaching assistantships - this advantage for U.S. citizens or permanent residents amounted to a female and male difference of only 3.38%. Males, on the other hand, were more likely to report research assistantships, trainingships, or internships as a primary source of support than were females. This difference amounted to 10.82% for U.S. citizens or permanent residents. Women also reported a higher reliance on personal funds than did men (female, male difference of 6.16% for U.S. citizens or permanent residents). Does the increased reliance of females on personal funds during the pursuit of the computer science doctorate discourage female participation in computer science?
  
• The majority of women and men who completed their doctorate degree between 1999 and 2000 remained in the United States. The majority of these individuals sought to find employment, rather than to continue study. Men reported a higher percentage of acquiring positions in industry, whereas women reported a higher percentage of acquiring positions in academia. What attracts women to careers in academia and men to careers in industry?

A natural question that emerges from this data is whether the identified differences are the result of choices that women make during their education or whether they are the result of subtle discrimination that occurs throughout their education? It is unlikely that any given study can capture the type of information that is necessary to answer a question this broad. The answer will most likely emerge slowly, from on-going systematic research that begins with the questions posed above.

Acknowledgments

The author would like to thank Susan Hill from the National Science Foundation for her assistance in acquiring the necessary reports to produce this paper, and for her assistance in the use of WebCASPAR.

Barbara Moskal is an Assistant Professor in the Department of Mathematical and Computer Sciences at the Colorado School of Mines.

References

Camp, T. (1997) "The Incredible Shrinking Pipeline", Communications of the ACM, 40(10), 103-110.

Hill, S. (2001a). Science and Engineering Doctorate Awards: 2000. Arlington, VA: National Science Foundation, Division of Science Resources Statistics, Science and Engineering Doctorate Awards: 2000, NSF 02-305.

Hill, S. (2001b). Science and Engineering Doctorate Awards: 1999. Arlington, VA: National Science Foundation, Division of Science Resources Studies, NSF 01-314.

Hill, S. (2000). Science and Engineering Doctorate Awards: 1998. Arlington, VA: National Science Foundation, Division of Science Resources Studies, NSF 00-304.

Hill, S. (1999). Science and Engineering Doctorate Awards: 1997. Arlington, VA: National Science Foundation, Division of Science Resources Studies, NSF 99-323.

Hill, S. (1997). Science and Engineering Doctorate Awards: 1996. Arlington, VA: National Science Foundation, Division of Science Resources Studies, NSF 97-329.

Moskal, B. M. (2002) "Female Computer Science Doctorates: What Does the Survey of Earned Doctorates Reveal?," ACM SIGCSE Bulletin - inroads, June 2002.

National Science Foundation (1996). Selected Data on Science and Engineering Doctorate Awards: 1995. Arlington, VA: NSF 96-303.

National Science Foundation (1995). Selected Data on Science and Engineering Doctorate Awards: 1994. Arlington, VA: NSF 95-337.

Sanderson, A. R. & Dugoni, B.L. (1998). Doctorate Recipients from United States Universities: Summary Report 1997. Chicago, Illinois: National Opinion research center at the University of Chicago.

Sanderson, A. R., Dugoni, B. L., Hoffer, T. B. & Myers, S.L. (2000). Doctorate Recipients from United States Universities: Summary Report 1999. National Opinion research center at the University of Chicago, Chicago Illinois.

Sanderson, A. R., Dugoni, B. L., Hoffer, T. B. & Selfa, L. (1999). Doctorate Recipients from United States Universities: Summary Report 1998. National Opinion research center at the University of Chicago, Chicago Illinois.

WebCASPAR (2001). "WebCASPAR: Your Virtual Bookshelf on Academic Science and Engineering." Arlington, VA: NSF [On-line]. Available at: http://caspar.nsf.gov/cgi-bin/WebIC.exe?template=nsf\srs\webcasp\start.wi.