Women in Computer Science 1/2: Beyond mere programming

Before 1995, the computer science department at Carnegie Mellon University, one of the top-ranked programs in the US, had low numbers when it came to admission and retention of female students: single-digit percentages for the admission, and of those few female students, many transferred out of the course. Starting in the early 2000s, the numbers changed dramatically: 40% of female students in the freshman year, and almost all of them then finished with a degree. What had happened?

The answer involves both Raj Reddy, Dean of the School of Computer Science from 1991 to 1999 (and himself a Turing laureate who attended the inaugural HLF in 2013) and Lenore Blum, a distinguished professor of computer science at Carnegie Mellon, who is attending this year’s HLF (with husband Manuel Blum, a Turing laureate). On Monday, Lenore gave a lecture on Alan Turing’s influence on “the other theory of computation” (namely numerical computations). When the session chair, in passing, mentioned her success in establishing a gender-balanced program at Carnegie Mellon, my curiosity was piqued. The next day, she was kind enough to tell me all about it, and it makes a fascinating story.

What, then, had happened? A number of different changes and, equally interesting, a number of things that hadn’t changed – in a combination that transformed the computer science culture at Carnegie Mellon, and should be of interest to anyone looking to establish gender-neutral university environments in the sciences, technology, engineering or math (STEM) subjects.

First of all, between 1995 and 1999, social scientist Jane Margolis and computer scientist Allan Fisher (then the associate dean for undergraduate studies) studied the interests, motivations and experiences of computer science majors at Carnegie-Mellon – where were the differences? What was keeping female students back? Their conclusions were published in book form as “Unlocking the Clubhouse: Women in Computing” in 2001.

The question whether or not a computer science program is welcoming to all genders (and, in a more general question, to under-represented minorities) has several facets. After all, if very few women apply to the program in the first place, that is part of the problem.

How to get women to apply?

When it came to applications, there was a happy coincidence that concerned the advanced placement programs (AP) in computer science – an opportunity for high school students to take college-level courses and examinations. Students with sufficiently high scores on the examinations are granted course credit by numerous colleges and universities.

At the time, in the mid-1990s, the AP program in computer science was heavily programming oriented. As language fashions in computer science changed, the AP program needed to change as well. In 1999, C++ replaced Pascal as the language of choice, only to be superseded by Java in 2003.

In the run-up to the 1999 transition, the US National Science Foundation offered funding for programs that would prepare computer science teachers for the change – and Allan Fisher realized that this was an opportunity to sensitize teachers to gender gap issues in computer science, as well! In a total of six week-long summer sessions, two each in ’96, ’97 and ’98, Fisher and his colleagues talked to the participating teachers not only about C++, but also about how to motivate girls to take up computer science. Whereas, in 1995, zero female students from those teachers’ schools had applied to Carnegie Mellon’s computer science program, in 1999, female students from those students accounted for 18% of the applications from prospective female students. When Blum, who had joined the Carnegie Mellon faculty in 1999, asked a number of those students why they applied, a frequent answer was that applying had been a shot in the dark for them, encouraged by their teacher(s). Evidently, the summer schools had affected a change.

How do we recruit future visionaries?

Carnegie Mellon is unusual among US universities in that students apply not for general admission, but for admission to a specific program of study. (Incidentally, that is how it commonly works at German universities, as well.) This gives the different schools that make up the university considerable influence over how to select their own students.

Before 1995, admission to the School of Computer Science had largely been based on existing programming skills at the time of application – a model that tended to amplify existing cultural differences, to whit: In a society where boys were mostly encouraged, and girls mostly discouraged, to tinker and program at an earlier age, selecting by programming skills will tend to select the boys, more specifically: the typical nerds.

Around 1995, Raj Reddy, Dean of the School of Computer Science since 1991, proposed that Carnegie Mellon should make an effort to aim higher. How, he asked, can we attract those people who will not only show technical excellence, but who are likely to turn out the world leaders and visionaries in computer science? Seeing the increased importance of computers in all areas of science, how can we attract the broad thinkers that will push the boundaries of computing in all the different possible contexts?

At about the same time, Allan Fisher lobbied for a more gender-balanced admission program, which would not invoke previous programming experience at all.

Balanced admission

As it turned out, those two goals dovetailed rather nicely. In response, the admissions office began to give more weight to non-academic factors, including leadership potential, achievements outside of academia, and a commitment to the wider community. The ideal image for a Carnegie Mellon computer science major shifted from the life-long computer tinkerer to someone who was a broad thinker, still good in math and science, but with broad interests, and not necessarily with previous programming experience. Admission officers began to look at the pool of applications as a whole, aiming at engineering a class that would be diverse and many-faceted – representative of the more general population that, after all, would be the users of new computer technology.

Among those admitted to the program, those that were new to programming would take an intensive semester of basic programming to get them up to speed; by the second year, everyone was to be supposed on an equal footing.

The new admission guidelines had positive effects – not only the number of women, but the total number of applications increased, from 1484 in 1995 to 3237 in 2001 (to over 6000 in 2014). During the same period, the percentage of female applicants that were actually admitted increased from 34% to 36%.

Blum is emphatic about one particular aspect that remained unchanged: The computer science program remained and remains one of the most challenging in the country – in particular the initial competitive and rigorous “boot camp”; any relaxation of standards would have endangered the status of the program as a whole. The new admission scheme only works because those who are admitted are up to that challenge. The difference is that those who are up to it now come from more diverse backgrounds than before.

No pink color computer science

There was one recommendation that came out of the research by Margolis and Fisher that the School of Computer Science did not put into practice. From their interviews, Margolis/Fisher had derived a fundamental difference in female and male students’ attitudes towards computers, with males predominantly seeing computers as an object of study and fascinations, while female saw computers as tools. Did this mean that a gender-balanced computer science curriculum should shift emphasis from the fundamentals to applications?

Blum refers to this solution as “pink color computer science”, and states that this was precisely what Carnegie Mellon did not do. Instead, when she came to the university in 1999, she took a different approach. There was something more important that needed to change.

Read more in Women in Computer Science 2/2: Changing the culture – coming soon to an HLF blog near you!