Scope of Research

After gaining approval from University of Georgia’s Office of Institutional Research under project title “Exploring Recruitment and Retention of Computer Science Majors,” the research team administered 527 surveys of computer science students and journalism students about their backgrounds, confidence, classroom experiences, future aspirations, perceptions of the CS department and attitudes toward technology.

The surveys were conducted in three different class levels and in five different sections. The students fit the following descriptions:

Gender:

• Male: 208

• Female: 63

Ethnicity:

• Caucasian: 164

• Asian, Asian American or Pacific Islander: 56

• Black or African American: 19

• Hispanic: 16

• Multiracial: 11

• Other: 4

• American Indian or Alaska Native: 1

Class level:

• First year: 74

• Second year: 77

• Third year: 58

• Fourth year: 58

Classes:

• 1301: 136

• 1730: 84

• 4720: 48

To gain further insight, we conducted classroom observations and interviewed 18 subjects — a mix of undergraduate students, graduate students and professors. Interviews were used to add narrative to some of the statistical results.

Survey Findings

Surveys were analyzed to determine the differences in responses between male and female. If the question related directly to experience with programming, class level was also taken into consideration. Women rated themselves lower in an overwhelming number of categories (47 in total). For questions that were also included on the journalism survey — for example, those regarding general confidence — the differences between male and female responses were more pronounced in computer science.

The largest differences are discussed in more detail below with accompanying recommendations (questions with most extreme difference are at the top):

• Confidence in programming abilities compared to peers in 1301

• Feeling need to constantly “catch up” in classes

• Preparedness for class compared to peers in 1730

• Confidence in programming abilities compared to peers overall

• Interest in CS compared to peers

• Confidence and comfort in using technology

• Confidence in current course of study

• Confidence in programming abilities in 1730

• Preparedness for class compared to peers overall

• Ability to succeed in computer science without losing balance in life

• Preparedness in 1301

• Similar career aspirations to peers

• Ability to find a job after graduation

• Peers valuing contribution in classroom

• Confidence in academic abilities overall

• Professor valuing contribution in classroom

• Becoming discouraged when working on CS projects

• Welcoming environment in CS classes

• Confidence in math abilities

Interviews

Interviews will be tabulated to tease out more data from responses. Here is a sampling of responses from the narrative of female computer science students at UGA:

“He looked at me and said: ‘Is that really what you want to do? You’re a girl.’”

“I’ve gotten: ‘Is it harder to program when you’re on your period?’ and ‘There’s girls on the Internet?’”

“I grew up in the boonies where if you get to 20-years-old without having a kid or getting married, you are doing pretty well. If I were to have changed my major then, my dad wouldn’t have agreed. It wouldn’t have been perceived as being girly.”

“In the CS Facebook group, someone posted a girl with an Apple ass from Reddit. That was put on the group with professors and no one took it down or commented. It was jarring.”

“It was a 60-70 person class. The second I walked in there, someone said ‘You’re the only girl in here.’”

“The professor would put stupid jokes on his slides. For a slide of a cash controller, he put a picture of a bride. He was just really kind of sexist, this old white dude. He didn’t notice. It was completely off his radar. I brought it up to another professor but she didn’t do anything.”

“He’s Chinese, and he’ll do really awesome coding, and everyone says, ‘Oh, it’s because you’re Chinese. That’s why you’re so good at doing this stuff.’ Our other friend, he’s also Chinese. He actually dropped out and went to MIS for his major instead, and we’re like, ‘But you’re Asian, like you have to be smart.’ It’s little things like that, which of course they think it’s all in good fun, but I understand that could hurt someone’s feelings.”

“I guess this is where the mean jokes come in; a lot of computer science women are associated with being really dorky, being sometimes lesbians, associated with being not pretty girls, and stuff like that. When there are pretty ones, people are like, ‘Oh my god, it’s a pretty computer science girl,’ and they kind of make a big deal out of it, or else they are kind of a little too harsh on the women that aren’t seen as attractive.”

Ties to other research

Many of my findings relate to research already done in the education and women’s studies fields. The findings are perhaps the most extreme in computer science, where unlike most STEM fields, women haven’t experienced a mentionable increase in enrollment in many years. From the surveys and interviews in computer science, we see an obvious manifestation of the following women’s studies findings:

The confidence gap: women are significantly less confident than men in their math and science skills, even when their achievements are comparable to those of their male peers.

Stereotype threat: Anxiety associated with a fear of confirming negative stereotypes about your social group.

Female undergraduate CS student: “There’s more pressure to excel. If you aren’t at the top, it brings all the female coders down.”

Imposter syndrome: You feel like an imposter, as if you don’t belong, and are unable to internalize your own success, which affects how you perceive your performance and abilities compared to your peers.

African American undergraduate student: “Because I guess it was a male dominated place and there is rarely a lot of African American so I did not feel, sometimes I feel like I didn’t belong here even though they were nice and helpful.  But in a crowd I felt like I didn’t really belong there.”

Attribution theory: Women are more likely to attribute failure to internal factors such as their intelligence and to attribute failure to external factors such as luck. Men do the opposite, which allows them to retain more confidence, even during times of failure.

Professor on catching female students before they drop out of the major: “Female students feel like, ‘Oh I’m just not good at this.’ They’ll just quietly go away. Male students will say, ‘The professor’s awful.’ And will blame it outwardly. I don’t know. It’s difficult to catch them.”

While the percentage of women has been steadily increasing over the last 40 years in almost all STEM fields, the percentage of women earning undergraduate computer science degrees has been dropping or stagnant over the last two decades (Gender Codes: Why Women Are Leaving Computing). The question arises, “What makes computer science different?” Largely, it’s the culture operating in computer science, which interview subjects labeled as “the boys club,” which only serves to aggravate the phenomenon women encounter in general education described above.

Those with the power to do something about the classroom environment in computer science are often those eager to push the blame to elementary, middle or high school. While it is true that interest can start early, there is plenty UGA and other computer science departments can do to better the environment for women and minorities in computer science. The department should concentrate on what can be done on its own turf instead of shifting the blame.

Recommendations

Provide multiple ways to enter the CS program

Male and female students become interested in computer science in fundamentally different ways. For male students, the most cited reasons they began CS were personal interest (42%), university classes (21%), encouragement from a family member that worked or was interested in CS (13%) or a class in middle school or high school (9%). For women, the breakdown is quite different — university classes (33%), encouragement from a family member that worked or was interested in CS (21%), personal interest (19%), class in middle school or high school (12%) or a friend that was interested in or worked in CS (9%).

From survey results, we know that women are less comfortable with technology and may have less experience with it. Creating three paths to enter the CS department — one for those who have extensive coding experience before the program, one for those with little coding experience, but comfort with technology and one for those who need more of a background in technology — would create a less pressured and intimidating environment in the early stages of the program and hopefully inspire more women already taking 1301 to continue with the program. One of the sections each semester should be concentrated on a topic that is multimedia-based and involve connections to other fields. The department would do well to incorporate guest speakers spanning a variety of career paths, genders and ethnicities to show students that they can find a place in computer science.

The classroom environment, especially in early classes, is intimidating for those that don’t have past computer science experience. These students are dropped in a room with other students (mostly male) who are already proficient coders and often aren’t shy about letting that be known.

One male student described it this way, “For me, that was a little bit intimidating at first because I didn’t understand a lot of what they were talking about. Even at the beginning of the semester I was considering maybe not doing this because I was intimidated by some of the people in the class. I don’t come from a background of computer science. A lot of the people in these classes have been coding for five and six years and they’ve made some apps for Android or Apple. That’s not something that I’ve ever even thought of doing. That seems so intangible to me.”

For a female student, a university class is more often than not her first experience with computer science. She experiences this intimidation differently because of stereotype threat and the other related issues described above. She may also have less family support for her decision to major in computer science, which causes her to question her fit even more, asking “Were they right?” Not one male student described a negative reaction from his family, whereas for female students interviewed, doubts from family on if they could handle the coursework were commonplace. Professors and students need to be aware of the privilege they receive as white males, their “invisible knapsack” as it were.

No male CS major has ever had to question if someone refused to work with him because of gender. This female CS student has, “There is one project that comes to mind in particular where the first week of class I showed up to a partner project. We were like we are going to meet on Wednesday, and we met on Wednesday, and he was like, ‘oh it’s already done’ and I am not sure if that’s him. I’m not sure because he didn’t trust me to do it. You never can tell what someone else’s motives are, but especially just coming in and having that at the forefront of my mind, like people are going to judge you because you are a woman. There is always a question there.”

Create a mentorship or tutoring program where upper-level CS students can buddy with new students. Provide office hour tutoring by upper-level female students and try to have one female TA available

Although women admit to feeling like they need to catch-up, feeling overwhelmed in the CS curriculum and not feeling prepared for their next class, they don’t go to professors’ office hours as much. While 14% of male CS students said they had never been to a professor’s office hours for their current CS class, 37% of women said they hadn’t been to office hours. Similarly, while 34% of male students go to office hours with a professor twice a month or more, only 5% of women go as much. Part of the reason for this is no doubt imposter syndrome.

One female student said the following of asking questions: “It’s just hard for me to admit that I don’t know something. Part of that is just my personality and part of that’s coming in and everyone already knowing everything. Not that many people know that I didn’t come in with any programming experience so I feel like if I am asking for help it’s going to be like, ‘Oh you don’t have a clue what you’re doing, you came in with no experience, and now you’re so behind.’ It’s harder to admit that you don’t know what you are doing in computer science than in any other subject for me. I just don’t want to be judged for it. There is a certain stereotype of computer science majors like the ultra nerd, so I don’t want anyone else to know that I don’t know what I am doing.”

Adding this type of a program would allow women to get help in a safer environment. Pair programming also helps with this. Professors should continue with pair programming projects and provide all possible checks to make sure students are programming together, at the same time and that one student is not completing the entire project without the other — something that was referenced often.  

Invite female speakers and speakers of color to campus and foster more opportunities for women to connect with one another outside the classroom

The survey results show that women students are more likely to feel they don’t have similar interests or career aspirations to their peers. This further confounds the notion that they don’t fit in. Students that are female and in a racial minority for computer science are hit with double jeopardy with this, where they see very few (if any) students or professors that look like them.

The examples of success in computer science are largely white men such as Bill Gates or Steve Jobs. Women and minorities need to be able to associate computer science with people they feel are similar to them. The department should include guest speakers who are women, of color or are pursuing career paths outside of software development at Microsoft to show all students how diverse computer science is outside of University of Georgia. Additionally, students should come away from the curriculum with some understanding of the history of computer science and women’s role in it over time. Although every student can name Bill Gates, few know Grace Hopper. This is a shame.

Many women also expressed interest in a group for women in computer science. An event like this every month or every other month that brought together female professors, TAs, graduate and undergraduate students would give students more cheerleaders and help them realize they are not alone in experiencing the feelings described above. ACM — the only existing group for computer science majors — isn’t well promoted. Increased promotion would bring in more students to this group.

Provide more support in 1730

As the department is already aware, 1730 is a sticking spot for many. Female students are likely to feel unprepared for the class. Male and female students alike recall it as the pivotal moment that caused them to stay in computer science or call it quits and switch majors. Having a “weed out” class is not helpful for the department in any respect. Past research has shown that women and minorities are the most likely to be “weeded out” in this type of a class. The department should consider changing the course structure or splitting the course in two to make it less overwhelming. The department should stop allowing students with no prior programming experience to take 1302 and 1730 as co-reqs. To help prepare students for 1730, extra review sessions, perhaps with women TAs should be added towards the beginning of the semester so that women feel more prepared in comparison to their peers.

Educate and engage students, teaching assistants and professors in conversations about diversity in computer science

Although male students had varying thoughts on diversity in computer science and admitted they were often bothered by jokes made in class, conversations such as this one were commonplace among the men I interviewed:

Interviewer: Why do you think more women don’t sign up for computer science classes?

Subject: I have no idea.

Interviewer: You have no idea?

Subject: No.

Interviewer: Have you ever thought about it?

Subject: I don’t know. I mean … is it the case that women don’t sign up for other science-type things, or is it just a computer science-type thing?

Interviewer: It’s just computer science.

Subject: Maybe they feel intimidated? I don’t see why they would.

There seems to be a divide in computer science. Those that still claim to be “blind” to race and gender can’t see the disparity that clearly exists around them. Those that notice the disparity, think about it often and have ideas on how to improve the situation. We need more of the second type. For those students (and professors) that have not noticed or have chosen not to address the glaring gender and racial disparities in computer science, whose responsibility is it to educate them? For professors, this is the department’s and the school’s responsibility. For students, this is the educator’s responsibility to address disparity in the curriculum. Students should not be able to graduate from an institute of higher learning in computer science without having a conversation about the racial and gender disparities in CS. Blind professors lead to blind students which lead to blind employees and blind bosses. The inclusion of diversity in the ethics class is certainly a start and this section should be continued and expanded upon.

Cut down on the “boy’s club” environment in some computer science classes by educating TAs and professors and encouraging them to speak up when they hear a sexist or racist joke. Make clear to students and professors that increasing diversity is a priority for the department and follow with actions. Provide more incentives for professors to promote diversity.

Professors and TAs must receive information from the department about minorities in computer science. They should know the data surrounding minorities in CS and understand how to translate the data to their own actions in the classroom. Awareness is the first step in cutting down on stories like these:

“There was this one other girl in the classes. Her lab partner was this idiot. He was the dumbest kid in the class, but he still couldn’t help but make the jokes of that girls aren’t good at computer science, and they were lab partners. The whole class she’s carrying him doing the work and he’s making the jokes about that she can’t.  He’s doing this loudly.  He’s like the class cut-up, and the professor is standing right up in the front not even noticing.  It doesn’t even register to him.”

On surveys, women scored the environment as less welcoming and thought peers and professors valued their contributions less than men did. Professors must be aware of research like this. It is unacceptable that someone could teach in computer science and not give a thought to the overwhelming gender disparity. At an obvious level, sexist jokes should be taken off all teaching materials such as slides and projects. Professors and TAs need training on appropriate ways to step-in and make clear that racist and sexist jokes are not going to be tolerated in UGA’s computer science classrooms. Additionally, professors need to be aware of the examples, teaching materials and career paths they talk about in class and to try to be as inclusive as possible. Although becoming a “hacker” is certainly a motivation for some students (mostly male), this is not the primary motivating factor for all students and shouldn’t be treated as such. The department needs to decide who’s responsibility it is to check for attitudes toward diversity and how these are manifesting themselves in classes. This group must include students.