A previous Communications of the ACM article [Camp 1997] described "the incredible shrinking pipeline" (ISP) for computer science (CS) women. The problem is seen clearly in two disturbing trends: (1) As students move through high school, college, graduate school, and academic positions in CS, the percentage of women drops . (2) In the years 1983-1996, the percentage of women obtaining bachelor degrees in CS went from a high of 37.1% to a low of 27.5%, a 25.9% decrease. Unfortunately, there is no evidence that these low percentages will rise.

A shrinking pipeline for women exists in computer science (CS), as it exists in other science and engineering fields. According to the most recent statistics available from the U.S. Department of Education (1995-96), the percentage of women who earned a bachelor's degree in CS was 27.5%, and the pipeline shrinks through the awarding of master's degrees (26.7%), and doctorate degrees (14.5%) [NCES 1998]. According to the most recent statistics available from the CRA Taulbee Survey (1997-98), a shrinking pipeline also exists in CS academic positions; 16.4% assistant professors, 11.7% associate professors, and only 7.6% full professors are women [Kozen and Morris 1999].

Focussing on a critical point in the pipeline, the percentage of degrees awarded to women at the bachelor's level, we find a disturbing trend that is unique to CS. The proportion of bachelor degrees awarded to women in all disciplines has increased almost every year for decades, to a high of 55.2% in 1995-96. This increasing percentage of women earning college degrees has had a positive effect on the percentage of women earning degrees in all science and engineering fields, except CS. In CS, the percentage of bachelor's degrees awarded to women decreased almost every year from 1983-84 (37.1%) to 1995-96 (27.5%). During the same period, the percentage of bachelor's degrees awarded to women in biological/life sciences, engineering, and physical sciences increased by 12.6%, 25.8%, and 30.4% respectively. We refer to the decreasing trend in CS as the incredible shrinking pipeline (ISP).

The ISP in CS is thoroughly discussed in [Camp 1997]. That article examines the percentage of CS degrees awarded to women from 1980 to 1994. Based on statistics gathered from a recent study, we speculate on what the future holds for women in computing through 2002.

Figure 1 plots the number of bachelor's degrees earned in CS by gender from 1970-71 to the most recent academic year available from the U.S. Department of Education [NCES 1998]. (See [Camp 1997] for a discussion on how the U.S. Department of Education classifies degrees earned in CS departments.) The figure illustrates that the number of degrees either increase or decrease for both men and women at the same time. The figure also illustrates that the increase or decrease in the total number of degrees is at a faster pace for women. In other words, the percentage of women earning CS degrees increases as the total number of CS degrees awarded increases and that the percentage of women earning CS degrees decreases as the total number of CS degrees awarded decreases . The ISP exists, as the decrease in the total number of CS degrees awarded to women far outweighs the decrease in the total number of CS degrees awarded to men (i.e., 55.7% decrease for women versus 35.1% decrease for men from 1985-86 to 1995-96).

Similar to the ISP problem in the CS departments surveyed by the U.S. Department of Education, the universities that are classified in the CRA (Computing Research Association) Forsythe List have also seen a decrease in the percentage of CS degrees awarded to women [Kozen and Morris 1999]. The CRA Forsythe List is a list of departments in the United States and Canada that grant a Ph.D. in computing. Since many universities and colleges have undergraduate and master's programs but do not award a Ph.D. degree, the data cited in the CRA Taulbee Surveys (which survey the universities in the CRA Forsythe list) says little about computer science as a whole. Figure 2 illustrates that the percentage of bachelor's degrees awarded to women in Ph.D.-granting CS departments is a lot less than the percentage of bachelor's degrees awarded to women in all CS departments (for example, 15.9% versus 27.5% for 1995-96). However, even within this subgroup of CS awarding universities, we still see a significant decrease in CS degrees awarded to women. Specifically, the percentage of women obtaining bachelor degrees in the CRA Forsythe list of universities went from 18.1% to 15.8% in three years, a 12.7% decrease. This decrease of women entering the information technology (IT) industry is of great concern for many reasons. As an example, the rapid growth of the computing industry has created a shortage of computer scientists (see [Arnheim 1997, Freeman and Aspray 1999] for statistics on the shortage today and [Keaton and Hamilton 1996] for speculations on the future). The ISP for women in CS has exasperated the IT shortage problem; if women were attracted to the computing disciplines at an equal pace as men, then even the tremendous shortages stated in the recent ITAA study could be filled [Freeman and Aspray 1999].

On the positive side, we should see an increase in the number of bachelor degrees awarded in CS in the near future. The number of new bachelor students enrolled in CS and CE (computer engineering) Ph.D.-granting departments increased 4% in Fall 1995 [Andrews 1996], 40% in Fall 1996 [Andrews 1997], and an additional 39% in Fall 1997 [Kozen and Zweben 1998]. While these statistics are taken from a small subset of all universities that offer a bachelor’s degree in CS, many non-Ph.D.-granting CS departments have reported a similar growth in their undergraduate enrollments [Walker et al., 1999]. One question, however, remains: how will this increase in the number of degrees awarded in CS affect the percentage of women recipients?

Due to a few volunteers obtained from the Systers list (www.systers.org) and a dedicated ACM-W intern, we have recently gathered statistics that answer this question. Figure 3 illustrates both the results obtained by the CRA Taulbee Survey (i.e., the percentage of bachelor's degrees earned by women in the universities classified in the CRA Forsythe list) and the results obtained from our study. In our study, we contacted each of the universities in the CRA Forsythe list to obtain the current percentages of women studying in the university's CS department, which gives us the expected Taulbee Survey percentages for 1998-2002. We received responses to our request from 80.2% of the universities in the CRA Forsythe list. Since the percentage of bachelor's degrees earned by women for 1997-98 from our study (16.6%) is comparable to the percentage of bachelor's degrees earned by women for 1997-98 from the CRA Taulbee Survey (15.8%), we are reasonably confident that our results are useful predictions.

Unfortunately, unlike the boom of CS students in the early 80’s, which led to a large increase in the percentage of women in CS (see Figure 1), our predictions show only a nominal increase in the percentage of women in CS followed by another decrease. This implies that the computing community cannot sit back and assume that as the numbers of students rises, the percentage of women students will automatically rise and the "ISP will take care of itself". We must take direct action to attract and retain more women to computing at all points in the pipeline (i.e., K-12, undergraduate, graduate, faculty and industry).

The good news is that the proportion of women earning bachelor's degrees in CS should increase slightly in the next few years at those universities that offer a Ph.D. degree in computing. We hope to see a corresponding increase in the percentage of bachelor degrees awarded in CS to women by non-Ph.D.-granting departments. The bad news is that the percentage of women earning bachelor's degrees in CS will continue to lag far behind their male colleagues in the near future.

ACKNOWLEDGMENTS

This work was supported in part by NSF Grants ANI-9996156 and EIA-9812016. We thank ACM for providing financial support for Vanessa Davies, the ACM-W intern that gathered the statistics presented in this paper.

ANDREWS, G. 1996. 1995 CRA Taulbee Survey. Computing Research News, vol. 8, no. 2, pp. 6-9, March 1996.

ANDREWS, G. 1997. 1996 CRA Taulbee Survey. Computing Research News, vol. 9, no. 2, pp. 5-9, March 1997.

ARNHEIM, L. 1997. IT Companies Heavily Courting CS Grads. Computing Research News, vol. 9, no. 3, pp. 1-7, May 1997.

CAMP, T. 1997. The Incredible Shrinking Pipeline. Communications of the ACM, vol. 40, no. 10, pp. 103-110, Oct. 1997.

FREEMAN, P. AND ASPRAY, W. 1999. The Supply of Information Technology Workers in the United States. CRA Publication, accessed Oct. 15^th, 1999.

KEATON, J. AND HAMILTON, S. 1996. Employment 2005: Boom or Bust for Computer Professionals? IEEE Computer, vol. 29, no. 5, pp. 87-98, May 1996.

KOZEN, D. AND MORRIS, J. 1999. 1997-98 CRA Taulbee Survey. Computing Research News, vol. 11, no. 2, pp. 4-9, March 1999.

KOZEN, D. AND ZWEBEN, S. 1998. 1996-97 CRA Taulbee Survey. Computing Research News, vol. 10, no. 3, pp. 4-8, March 1998.

NCES. 1998. Digest of Education Statistics. National Center for Education Statistics, U.S. Department of Education, Washington, D.C.

WALKER, H., MYERS, J. P., ZWEBEN, S., TUCKER, A.B., AND BRAUGHT, G. 1999. The Crisis in Academic Hiring in Computer Science. Proceedings of the 30th SIGCSE Technical Symposium on Computer Science Education, pp. 362-363, March 1999.