The Anatomy of Interest:

Women in Undergraduate Computer Science

Jane Margolis, Allan Fisher and Faye Miller

School of Computer Science

Carnegie Mellon University

5000 Forbes Ave.

Pittsburgh, Pennsylvania 15213

(412) 268-6630

Lily was a first-year undergraduate computer science major who entered one of the top computer science departments in the country with a great deal of enthusiasm for the subject. Her interest was first sparked in high school, when she took an advanced placement computer science course at the suggestion of her guidance counselor. "As soon as I started taking that course in programming, I realized I loved it...I absolutely loved it," she tells us in her first interview. Her enjoyment of a summer programming job solidified her decision to major in CS. She enjoyed "the challenges the programmer faces," and found the problem-solving to be fun.

Yet by the end of her second semester, she has decided to transfer to the English department. Her enthusiasm for computer science is extinguished. She says, "In high school, when I’d go home from class, I would be like ‘Oh, let’s program a little.’ But, now I am just like, ‘Let’s not bother.’" Struggling with the course work, perceiving her peers (mostly male) as doing much better with much less effort, feeling a misfit between herself and a cultural norm that associates success with an all-consuming love of computing, she questions whether she belongs in the department. Several semesters after leaving computer science, Lily describes her disappointment in having transferred out. It is not that she is unhappy in English. She loves the humanities. But she remembers how much she loved programming, had wanted to major in computer science, and feels dismayed with how her interest has been extinguished.

Lily’s experience is not unique. Her story highlights a key struggle experienced by many women studying computer science at the college level. Once enthusiastic about the field, their interest dissipates. Departments have witnessed many of these students leave computer science, often around their Sophomore year-- the time when many students, male and female, regardless of major do their "switching." It is then too convenient for all concerned to conclude that Lily, and others like her--those "appropriate switchers"¹ -- have found their intellectual interest and passion elsewhere. The professors may even be thankful that these students made this self-discovery now rather than later. In computer science, a higher proportion of students who transfer to other majors are women.²

In this paper we focus on the process by which women students who enter with high enthusiasm and interest in computing quickly lose faith in their ability and their interest in the subject. We believe it is critically important to gather carefully rendered accounts of how these women students lose interest in computer science, particularly in light of the commonly-held belief that girls and women are not as "intrinsically interested" as males in technology. We look at the ways that interest in a subject area is influenced by factors beyond an individual’s intellectual preference for an abstract body of knowledge. We have found women’s departing statements that they are "just not interested" to be a misleading endpoint to a complex process we’ve seen over time, involving the interplay of gender-biased norms of interest and eroded confidence. We look at the mechanisms that promote a male focus of interest as the standard for success, while casting doubt upon women’s interest and ability in computing.

At stake in the experience of women as undergraduates is the makeup of the technology-creating population. While girls and women may be using the Internet for communication and the web for information retrieval, it is predominantly men (the majority in the U.S. being white and Asian-American) who are programming the computers, designing and fixing the systems, and inventing the technology that will affect all aspects of our lives (National Science Board 1998). Only 15-20% of undergraduate computer science majors at leading U.S. departments and only 17% of the high school computer science Advanced Placement test takers are female (College Board 1998; Kozen & Zweben 1998). The under-representation of women and minorities among the creators of information technology has serious consequences, not only for those individuals whose potential goes unrealized, but also for a society increasingly shaped by that technology.

Our research team is unique in that it is interdisciplinary, made up of researchers from computer science and education/women’s studies. The heart of our research is interviews with computer science students themselves. Students are initially interviewed about their prior experience with computing at home and at school; then, once per semester, they are interviewed about experiences in the program, feelings and attitudes about studying computer science. Additional data is gathered through surveys, classroom observations, interviews with faculty, a small journal writing project, and monitoring of electronic communication and forums. Our sample consists of the majority of female computer science majors at Carnegie Mellon, and a comparable sample of male majors. We have interviewed 51 female and 46 male computer science majors, most multiple times, for a total of 210 interviews with 97 computer science majors. The majority of the students are European-Americans and Asian.³ At the conclusion of this year we will have followed samples from three incoming classes through four to two years of study.

The women who are admitted to the Carnegie Mellon School of Computer Science enter with high enthusiasm and interest. We’ve interviewed a student who was the president of her school’s computer club, another who took great pride in being the "computer genius" in her family. Almost every woman in our sample came to computer science because of an interest and strength in math and science; they also enjoy problem-solving, puzzles and logical thinking tasks. They take pleasure in programming. As one student tells us: "I have always been a problem-solving person and it just really gives me a rush when you are working and working on a problem and finally it just works!" A number of the women express delight in the sense of mastery of figuring out the program. Women talk about the pleasure in "systematic thinking," as well as the creative aspects of programming.

While the women students enjoy the "rush in having my program run," the context and connections of computing to other arenas often makes the study of computer science meaningful. Forty-four percent of the women students in our study (as compared to 9% of the men students) contextualize their interest in computers in other arenas such as medicine, space, or the arts. These women emphasize the importance of integrating computing with people and other arenas. As one student tells us:

A first-year student told us how much she enjoyed science fiction when she was young. She remembered reading about a robot that was "more like a tool; it wasn’t something that would take over a place, but it was a machine that would help out." She wants to design this type of intelligent machine. Inspired by a lecture about a robot car (in which the lecturer explained the utility of the car by describing the number of accidents and deaths caused by human error), she explained how this people-oriented purpose for computers is what resonated with her desire to connect computer science to real world problems:

Our finding that women are concerned with the usefulness of computers is concordant with other research from the field (Honey 1994; Martin 1992; Schofield 1995). While women’s interest and enjoyment of computing itself is often a primary reason for majoring in computer science, most also tend to have a multitude of practical and related reasons for choosing their major, such as the desire for secure employment, the versatility of computer science, encouragement from others, and the exciting, changing field.

Despite the fact that most of the women students educated in the U.S. had computers in their homes growing up and enjoyed using them, their interest in computing was usually one among several and developed more gradually than it does for the male students. As the women describe their enjoyment of programming, most compare themselves to a male member of their family, and note that he was the one who was "really into computers." More female students in our sample report watching while a male family member (father or brother) played games, tinkered, or took the computer apart. As one female student tells us: "We got a computer when I was...I don’t remember how old I was, but my brother just totally took to it and that’s when I wondered, ‘What is this thing?’" There are no subsequent references to her own self-exploration. Another student, in describing her computer experience, also draws a clear contrast between herself and her brother:

This fascination with the computer often leads to self-initiated exploration and learning about various aspects of computing, including programming. Computer games are predominately designed with boys’ interests in mind, and have much to do with this (Cassell & Jenkins 1998). One student told us how games were a key source of his motivation to learn programming:

This description of the pull to "figure out how it works" is something prevalent in the male interviews and very rare in the female interviews. It positions the male students in a very active relationship to the machine. As young boys, males step right into the driver’s seat, leaping from the outside to the inside. Many male students report programming to be a source of extracurricular pleasure, having done it since they were young (38% of the males, compared to 10% of the females in our sample). Computing emerges as an integral part of their lives.

Many women do enter undergraduate computer science with less computing experience than their male peers (Klawe and Leveson 1995; Teague and Clarke 1996). To assess gender differences in prior programming experience of our 1995 interviewees, we conducted a survey of incoming students, and found that 40% of males respondents from the Carnegie Mellon first-year class passed the advanced placement computer science exam, thereby placing out of the Carnegie Mellon introductory level programming class, while none of the first-year women placed out. Men were also familiar with more programming languages than females, and were much more likely than women to report having an "expert" level of knowledge of a programming language prior to Carnegie Mellon. Despite this gender-computing experience gap, it is important to note that we have found no correlation between prior experience and success; some of the most successful students have entered the program with very little computing experience. Yet when women students encounter difficulty with course work, these comparisons between their experience and that of their mostly male peers can become the kiss of death. As one woman said:

Clearly, a wide range of psychological, socialization, and pedagogical factors contribute to women perceiving themselves as "picking up ideas slower" and their male students as "so good without even trying." For instance, while males often have more experience, there is also the likelihood that male students are inflating how much they know, or are overly confident, while women have been shown to underestimate their abilities (Sax 1995; Lundeberg, Fox & Puncochar 1994). Much other research and our own observations point both to the fact that men tend to exaggerate their achievements more than women, and that men are more likely to attribute their successes and failures in a self-affirming way than are women (Eccles 1992; Dweck 1986; Licht and Shapiro 1982). Strenta found that "even with the same grades in science and other courses, women in science were less confident of their abilities and more depressed about their academic performance than men" (Strenta et al. 1994, 543). We see both of these effects clearly in our interviews, and we suspect that a greater knowledge of the prevalence of these trends would help students to insulate themselves. And although prior computing experience is not a criterion for admission to the program, many students find that introductory courses do assume a certain amount of technical familiarity that many novices do not have, and seem to assume a background that is far more typical for men than women in the program. To create gender equity at the undergraduate level, computer science programs must address the question of the unlevel playing field in terms of prior experience. Other research studies of women in the sciences and engineering (Brainard and Carlin 1997; Seymour and Hewitt 1997), have found a similar pattern of the unraveling of undergraduate women’s confidence during their first year.

In their article "Encountering an Alien Culture," author Sproull describes the college computing culture as an "adolescent one" in which:

An intense focus and love of the computer is part of being "in synch" with the culture (Kidder 1981; Hafner & Lyon 1996; Levy 1984).

A female junior student who was very involved in the Internet before coming to Carnegie Mellon and who has always regarded herself as a math-science person, after several semesters doesn’t think that computer science is for her because, "It’s not my passion like everyone else. They’re all, like, really into it." In her particular case her boyfriend is "really into robotics," planning on going to graduate school and becoming a professor, but she sees herself as different. When the interviewer asks her to "talk to me a little more about the reasons you are thinking that computer science isn’t for you," she says:

Comparing herself to peers who "dream in code," she feels she is not meant to be in computer science. Another woman feels that her male peers "have a motivation that’s deeper than I do" and "if they love programming that much" she wonders if she should be in the major. She continues:

A woman student who ended up transferring out of Computer Science to Information

Systems describes how she was not interested when her classmates would talk about the "latest machinery on the market or how much RAM space a computer has and how they have to upgrade it and what new software they want (blah, blah, blah)." She concluded:

Twenty percent of the female computer science majors we interviewed have questioned if they belong in computer science because they feel they do not share the same intensity of focus and interest they see in their male peers. Women describe wanting to talk about other things besides computers, feeling estranged from those who are myopically focused on a machine.

Sheila Tobias, in her study of women in science, argues that the "boy wonder icon," the association of male traits with success in science, is central to the broad assumptions of who does (and who does not) become a scientist:

The computing culture, in particular, reflects the history and legacy of actual "boy wonders." The widely-held expectation that there is one way (the male way) to come to science shapes the expectations and assumptions of parents, teachers, and students themselves. As women begin their undergraduate computer science education, and perceive that men go through the program with more experience, less effort and more interest, too many women conclude they do have to look "like him" in order to succeed. Women begin to question their ability and their interest -- whether it is "enough."

These male-biased expectations of who will succeed affect teachers as well. A particularly clear illustration of this is captured in an e-mail exchange between two high school computer science teachers who had participated in a Carnegie Mellon summer program for advanced placement computer science teachers [http://www.cs.cmu.edu/6apt]. A male teacher began the discussion about the low numbers of female students in his computer science classes:

A female teacher responded that many of her students make the same observation as above, having the idea that "staying up all night programming is a sign of love for computer science, and not doing so is a sign that one doesn’t love it." But she disagrees. She talks about taking her first programming course in college and how she "fell in love with it." She said it was "organized, logical and yes, fun." However, she did not stay up all night doing it. She says she "did not even spend a majority of my time programming" and did not program on her own, coming up with games, entertainment, etc. She said she enjoyed the programming assignments "immensely," enjoyed the challenge and especially enjoyed the "practical problems." But then she adds:

We concur that there are multiple ways of being interested in a subject, and that valuable contributions to the field come from people with different sets of attachment to computers. Students can love computer science without being "like him."

Yet despite these feelings of difference, we find that male students report less distress, are less affected by the perceived difference between themselves and their peers, and leave the major in smaller proportion; and despite resistance to total absorption in computing, they do not question their ability to become computer scientists if they choose to do so, and they seldom feel the need to conform or change to match this norm. They do not speak of being discouraged by others who are passionately involved with computers and thereby know more. And from the men with less experience, we do not hear anguish over whether they "have what it takes." They describe how they perceive themselves as different from the dominant image, but this does not excessively discourage them or immobilize them with self-doubt.

We believe the reason why female students feel more distress over this difference from the stereotype than non-hacker males, is that the computer science culture assumes that men will succeed. Success is linked to a stereotype based on a common male pattern of desires, interests, and attachments to computing. Hence it bolsters men’s confidence and sense of belonging. This same culture does not assume (often accurately) that women conform; hence they enjoy no default expectation of success, and their interests and attachments to computing may be regarded as deviant from the norm, and less serious than those of the male students. This, combined with a vast array of gender socialization factors, chips away at women’s sense of confidence and belonging in the field. Men who face difficulties with coursework do not struggle under the additional burden of the presumption that they are somehow inferior by virtue of their gender; nor do they have the pressure of feeling they are representative of their gender. One woman student hit the nail on the head:

Seymour and Hewitt (1997), in their multi-university study of students leaving math and science majors, concluded that "intrinsic interest" in the subject was the number one factor in helping students persist in the face of various obstacles, including poor teaching, lack of support, and a chilly climate. The importance of "intrinsic interest" makes intuitive sense. Yet our research identifies a "chicken and egg" problem that affects a female student as she thinks about her interest in computer science: is difficulty with course work due to lack of experience? or is lack of experience due to lack of interest? or is lack of interest due to lack of confidence? or is lack of confidence due to lack of interest? It is hard to disentangle one influence from the other, to know how they interact. Women’s interest (or lack of interest) may not be as "intrinsic" as it feels, for interest is continuously encouraged or extinguished, and defined, by cultural norms, external factors and internal responses. Far from being a static, immutable given, a combination of gender socialization, cultural artifacts of the field (a male-dominated history, culture, educational practices, peer interactions), and individual psychology all construct and/or undermine many women students’ sense of their own "intrinsic interest."

The words of one of the women we interviewed, a junior at the time, capture well what we have identified as a nexus of confidence and interest that entangles many of the women computer science undergraduates:

A woman student who has been playing with computers since she was four, coding since she was five or six, was on the high school programming team, won prizes etc. tells us how comments from male peers shook her confidence, which then diminished her interest in programming:

Research shows that both males and females believe that males are better than females at computing (Clarke 1992; Spertus 1991). These low expectations for female students becomes part of a biased social environment, often manifested in the form of peer-to-peer interactions. The following post by a male student on a Carnegie Mellon computer science undergraduate electronic bulletin board on why there are so few women in the department:

Comments from male peers, which may appear incidental or random, accumulate to make women feel undervalued, and unwelcome (Valian 1998; Spertus 1991). Women report being faced with comments like "You are a computer science major and you don’t knowthat?" if they don’t know a technical detail or ask a question. The most damaging comment, and perhaps the most telling, is one that one quarter of the women we have interviewed mention having heard: that they "got in" to the computer science department only because of their gender:

While most of the women say that the majority of their male peers are nice and helpful (their perception is that just a few are "bad eggs"), it is clear that a biased belief system affects peer interaction. Research from other universities shows a similar pattern. The MIT Artificial Intelligence Laboratory report "Why Are There So Few Female Computer Scientists?" concludes that comments and behaviors such as those cited above are "the symptom of a more fundamental problem: lower expectations for females" (Spertus 1991, 14). While there is no way for us to know how much effect this biased social environment has on individual women, research across fields, studying what matters in college for students, concludes that "the students peer group is the single most potent source of influence on growth and development during the undergraduate years" (Astin 1993).

Psychologist Claude Steele’s examination of the experiences of African Americans in higher education and women in traditionally male fields suggests that when one is in a situation in which a negative stereotype about one’s group applies, one is fearful of confirming the stereotype. This can lead to poorer performance (as in test scores and work performance); in turn, this turns into disidentification with the field. Disidentification with a field offers "the retreat of not caring about the domain in relation to the self. But, as it protects in this way, it can undermine sustained motivation in the domain" (Steele 1997, 614). This is one way of understanding students’ statements that "computer science just doesn’t interest me," as an endpoint of experiencing difficulty with course work and a drop of confidence. Steele believes that "one must perceive good prospects in the domain, that is that one has the interests, skills, resources, and opportunities to prosper there, as well as that one belongs there, in the sense of being accepted and valued in the domain" (Steele 1997, 613) in order to persist, and maintain one’s interest and engagement.

We believe that this social environment, and the biased assumption that females are not likely as males to be capable computer scientists, is made ever more potent by the following characteristic of the field: the range of student knowledge and skills appears to have no upper limit. Because of adolescent male interest (and energy) which often is directed into investigating of everything about computers, and the availability of vast informal learning opportunities, many undergraduate students (mostly male) enter the program with what seems like a limitless amount of knowledge. In some cases, the "whizzes" and "experts" are the students, as well as the faculty.

A fascinating counterpoint to the stories above is the stories of the international women, many of whom enter the undergraduate major with the least experience and "intrinsic interest" in computer science, yet persist and succeed in the program.⁴ Kanitha, from Thailand, told us:

These students often go on to employment in the computing field or to graduate school in computer science. Furthermore, their persistence often leads to an intellectual pleasure in computer science that they often did not have when they began. A second-semester student from China speaks of how her attitude toward being in computer science has changed:

The stories of these women who enter with little intrinsic interest and experience in computers challenge the stereotype of who can do computer science. While the international female students enter with the least computing experience and least intrinsic interest, as their hard work takes effect, the mystery of computer science begins to metamorphose into mastery. They then often begin to find intellectual pleasure in the field. It becomes an "acquired taste." A Malaysian student told us of the eventual satisfaction she got from sticking it out:

Rather than epiphany moments of falling in love with computing the minute they touch the keyboard and the computer responds, as described by male students, the stories from international women present a different trajectory to becoming a computer scientist, in which their interest in computers evolves over a longer time. But it takes a completely different cultural orientation, and different set of motivations, to overcome the biases in the program that they also face.

Our data suggests that many of U.S. women who persist are ones who have found a way to get the grades they are satisfied with, are able to reconcile a different relationship to computing, and maintain confidence in the face of more experienced peers. They also conclude that they don’t have to measure themselves against the male norm of interest. For instance, a female student who plans on using computing to improve tutoring programs for children, described how she has refused to conform to the image of the myopically focused "computer geek" who "hacks for hacking’s sake." And since she is "getting really good grades without changing myself" she is ever more confident that she can remain in the major and be herself. When the interviewer asks her if she feels a need to conform to the stereotype, she answers "I refuse to." Then she adds:

While there is a large inherited component to the gender gap in higher education computer science, we believe that there is much that higher education can do to protect and engage the interest of women students in computing. Our understanding of the problem, some initial intervention experiences and other research suggest key areas for intervention: culture, curriculum, and confidence. An important theme throughout each one of these is that women must not be expected to model themselves after the stereotypical male computer science student.

Some of the elements of a more contextual approach include early experiences that situate the technology in realistic settings; curricula that build upon the connections between computer science and other disciplines; and diverse problems and teaching methods that appeal to a broad variety of preferences and styles. At Carnegie Mellon, some of the approaches that we have adopted include:

• Interdisciplinary courses that bring students of diverse backgrounds together to work on multifaceted problems;
• An undergraduate concentration in human computer interaction;
• A course that engages students with non-profit groups in the local community, applying their skills to community issues.

All of these efforts provide additional paths for students to pursue in addition to the traditional, technically focused path, and more remains to be done. Sue Rosser, in her book Female Friendly Science, argues that "insuring science and technology are considered in their social context may be the most important change that can be made in science teaching for all people, both male and female" (Rosser 1990, 72).

Claude Steele also emphasizes the importance of positive student/faculty relationships. Steele reasons that the student "must feel valued by the teacher for his or her potential and as a person. Among the more fortunate in society, this relationship is often taken for granted." Building this relationship of respect between teacher and student is "the first order of business--at all levels of school. No tactic of instruction, no matter how ingenious, can succeed without it" (Steele 1992, 77).

From our interviews with female undergraduate computer science students, we concur with the importance of faculty relationships and support for women students. Women students describe how they got turned on to computer science and began to consider a major in it because of a high school programming class they enjoyed and were good at, and a teacher who worked with them and encouraged them. They also talk about the importance of the support, advice, and guidance they get from the faculty members at Carnegie Mellon who teach and advise first-year students.

A more difficult feature of the confidence problem is the role played by students’ self-evaluations. A partly intellectual, partly cultural, agenda that we have considered pursuing is to raise awareness of the issues that affect women students’ self-assessment and confidence. One approach would be to provide students with some information on cognitive psychology, hoping this will help dispel some of the myths about effortless learning and innate ability that seem to have a corrosive effect on self-confidence (Dweck 1986). Departments also need to find ways to reduce the occurrences of comments such as, "You are only here because you are a girl." Students must be educated about admissions policies that show that this is not so, and also understand that the institution considers taunting such as this to be unfriendly and hostile to students’ learning environments.

In this paper we show that behind the commonplace observation that "women are not interested in computer science" lies a complex of influences. While some people believe that it is the inherent nature of computing itself that turns women away, we have documented social and cultural expectations within the field that discourage girls and women. Our close-up look at the experiences of women in undergraduate computer science, who were once very enthusiastic, reveals some of the pernicious ways in which male behavior and interest become the standards for "the right fit" and success. Fast-forwarding the tape from their initial interview where students are excited about learning computer science to the following two to three interviews, we hear too many women students feel like they are "drowning," believing they lack the interest to keep afloat. Trying to find their place in a culture that challenges whether they are "really into it" and a curriculum that assumes their learning will occur in the same sequence and timing as their male peers, too many women conclude that they "just aren’t interested." Interest is a precious thing to have. We believe its loss is unnecessary and unjust.