Maria Klawe, President, Harvey Mudd College
In the 40 years since the publication of Social Issues in Computing, new awareness of a significant issue has arisen: the lack of women in computing and what can and should be done about it.
While women’s participation in the other STEM disciplines has risen over the past three decades, participation in computer science has decreased. In Canada and the U.S., the percentage of women who graduate from college with a degree in computer science is at a 30-year low. In 2012, only 13 percent of undergraduate computer science majors were female (Computing Research Association, 2012 Taulbee Survey, www.cra.org/resources/taulbee).
Women represented nearly a third of those receiving bachelor’s degrees in CS during the technology boom in the 1980s. Yet a decade later, the percentage of women majors had dropped to 20 percent, and for the past 10 years, the national average has hovered around 12-14 percent.
The dotcom crash in 2001 caused both male and female students to lose interest in the major, but the decrease over the last decade was nearly twice as sharp for women. One report found that the number of males receiving CS bachelor’s degrees at research institutions fell 35 percent (from 10,903 in 2001 to 7,039 in 2009). The corresponding number of females plummeted 67 percent (from 2,679 to 892). Only now are we starting to see a turnaround as the U.S. enters a new tech boom, but female CS participation still greatly lags that of males.
The current demand from industry for software developers, computer scientists and computer engineers far outpaces the supply of CS graduates. This shortfall impedes technological and economic advancement. We cannot meet the needs of industry if we are drawing from half the population. We also cannot develop the best, most creative solutions when teams are homogenous. Clearly, there is a need to ignite women’s interest in CS, particularly at the undergraduate level.
Why so few women?
With major technology companies and startups competing intensely for talented college graduates in computer science, why do so few women enter the field? Research shows that young women are reluctant to study computer science for three reasons: 1) Young women think computer science is boring; 2) Young women think they won’t be good at computer science; and, 3) Young women think they will not feel comfortable in the CS culture–they view computer scientists as nerdy people with poor social skills.
Successful efforts to increasing participation
Fortunately, colleges and universities that have made a serious commitment to increasing female participation in CS have had substantial success—starting in the mid-1990s with Carnegie Mellon University and the University of British Columbia, and, more recently, Harvey Mudd College. Over a four-year period, Harvey Mudd quadrupled the percentage of CS majors who were female from 10 percent in 2006 to 42 percent in 2010. Since then, that percentage has ranged between 35 to 45 percent. The CS department accomplished this by implementing three innovative practices: a redesigned introductory course, early research opportunities during the first few summers, and sponsored trips to the Grace Hopper Celebration of Women in Computing. These changes plus the recent intense demand for CS graduates also resulted in a dramatic increase in the total number of CS majors at Harvey Mudd, which have roughly tripled from 2006 to 2013. Thus the actual number of female CS majors graduating from Harvey Mudd each year has increased by close to a factor of ten over that period. While institutions differ in many respects, most of these approaches can be replicated or adapted by institutions seeking to encourage women to study CS.
The redesigned intro course, CS 5
To spark interest in CS—and demonstrate that it is anything but boring—Harvey Mudd’s computer science faculty reframed the introductory course, CS 5, from “learn to program in Java” to “creative problem solving in science and engineering using computational approaches.” The new course covers the same computer science concepts at the same level of rigor as before, and students do even more programming, though in Python rather than Java. Students find Python a more flexible and forgiving language than Java, and unlike some other “easy to learn” languages, Python is popular in industry. Students are given a choice of problems for each homework assignment in contexts such as epidemiology, robotics or physics. Both male and female students report that they are excited—and often surprised—to discover that CS is a much more fascinating and rewarding discipline than they previously thought. As a result of these changes, CS 5 went from being Mudd students’ least favorite required first-semester course to the most popular. Moreover its popularity spread to students at the other Claremont Colleges. This semester more than half of the students in CS 5 were from the other colleges.
To increase students’ confidence, CS 5 is divided into sections according to their prior computing experience. Students with little or no programming experience are placed in CS 5 Gold while students with substantial programming experience are in CS 5 Black (Harvey Mudd’s school colors are black and gold). Students with even more programming experience are placed in CS 42, a course that combines the material in CS 5 with the next CS course, CS 60. In some years Harvey Mudd also offers CS 5 Green, which covers all the material in a biology context. In all sections, instructors deliberately work to eliminate the “macho” effect—where a few students with more experience inadvertently intimidate others by showing off their knowledge—and discourage less-experienced, but equally talented, classmates. We have found a private conversation along the lines of “I’m delighted to have such a knowledgeable student in my class, but some of the other students may find your knowledge intimidating, so I’d like to have our conversations on a one-on-one basis rather than in class” very effective. Eliminating the “macho” effect in our introductory courses has significantly improved the culture in all CS courses at Harvey Mudd and resulted in a more supportive learning environment for all students. The grouping by prior experience has created a confidence-boosting atmosphere, especially for beginners, who are disproportionately women and students of color.
Early research opportunities
A number of studies have shown that research experiences for undergraduate women increase retention in STEM fields and the likelihood they will attend graduate school. From 2007 to 2010 a grant allowed us to offer summer research experiences to about 10 female students at the end of their first year. Faculty created research projects suitable for students who had completed only one or two CS courses; these projects allowed first-years the chance to immediately apply their knowledge, boost their confidence and deepen their interest in the discipline. Students embraced this opportunity to engage in 10 weeks of intensive, challenging summer research on projects such as artificial intelligence, robotics and educational video games. They discovered they could do CS research, do it well and enjoy doing it. This program helped increase the number of female CS majors during the years of transitioning from 10 percent to about 40 percent. By the time we achieved critical mass in terms of the number of women in all CS courses, we no longer found it necessary to continue this program, though we still offer summer research experiences to many students, male and female, at the end of their first year.
A welcoming CS culture
To increase female students’ sense of belonging in the technology field, in 2006 Harvey Mudd began taking a large cohort of first-year female students to the Grace Hopper Celebration of Women in Computing. At Hopper, students see the variety of jobs available within the discipline and meet successful role models at all career stages, as well as experience an effervescent and welcoming culture. The conference has proved to be a powerful tool in encouraging young women to take more computer science classes and ultimately major in computer science.
The path to choosing to major in computer science
The redesigned CS 5 introductory course and the Hopper conference work together to encourage female students to take the next course in the CS sequence, CS 60. By ensuring that the CS 60 experience is also interesting and enjoyable, we motivate many to take the third course, CS 70. For most of our female students, it is during or after CS 70 that they decide to make CS their major.
Transferability to other institutions
A National Science Foundation grant (CPATH-2) for $800,000 allowed us to disseminate our highly successful CS 5 curriculum and share our approaches with other institutions, many of which are now teaching the course in its entirety or adapting it with great results. Of course Harvey Mudd has two advantages that are not present in every institution. First, all students must take a CS course in their first semester. Second, students do not have to choose their major until the end of their second year, which gives them time to try out several CS courses before choosing a major. Still, every institution can make its introductory course fun and non-intimidating. CS 5 is not more expensive or challenging to teach than other introductory courses. Eliminating the “macho” effect takes a very modest amount of extra time by instructors and has a huge payoff. Taking students to Hopper costs about $750 per student, depending on the location. There are now several regional versions of Hopper, making attendance even more affordable. Finally, many foundations provide support for summer research for undergraduates. Another approach that has worked well at UBC is to encourage double majors in disciplines with large percentages of females such as biology, chemistry, psychology or statistics. The truth is that every CS department that makes a serious extended commitment to increasing women’s participation in CS can make substantial progress.
The world of computing–and the world in general–will benefit.
For more information about Harvey Mudd’s effective practices to increase women’s participation in computing careers, please see:
Evaluating a Breadth-First CS 1 for Scientists. Dodds, Libeskind-Hadas, Alvarado, Kuenning. In Proceedings of SIGCSE 2008 http://www.cs.hmc.edu/~dodds/sigcse08FinalWNames.pdf
Women in CS: An Evaluation of Three Promising Practices. Alvarado and Dodds. In Proceedings of SIGCSE 2010. www.cs.hmc.edu/~alvarado/papers/fp068-alvarado.pdf
HMC CS 5 course website: http://www.cs.hmc.edu/twiki/bin/view/ModularCS1/
For more information on national statistics and issues surrounding women in CS:
Women in Computing–Take 2. Klawe, Whitney and Simard. In Communications of the ACM 2009. http://cacm.acm.org/magazines/2009/2/19326-women-in-computing-take-2/fulltext
Computing Research Association’s Taulbee Report: http://www.cra.org/resources/taulbee/
Watch a short video of one of Harvey Mudd’s engaging CS summer research projects: http://www.youtube.com/harveymuddcollege#p/u/11/o_O2o2ly-34Maria Klawe began her tenure as Harvey Mudd College’s first female president in 2006. Prior to joining HMC, she served as dean of engineering and professor of computer science at Princeton University. Klawe joined Princeton from the University of British Columbia where she served in various roles from 1988 to 2002. Prior to UBC, Klawe spent eight years with IBM Research in California and two years at the University of Toronto. She received her Ph.D. (1977) and B.Sc. (1973) in mathematics from the University of Alberta. Klawe is a member of the board of Microsoft Corporation, Broadcom Corporation and the nonprofit Math for America, a fellow of the American Academy of Arts & Sciences, a trustee for the Mathematical Sciences Research Institute in Berkeley and a member of both the Stanford Engineering Advisory Council and the Advisory Council for the Computer Science Teachers Association.