Retaining Female Students in Computer Science
This research done for Southern Oregon University kickstarted my passion for supporting marginalized groups in computer science. It also showed me how much I enjoy research with an end goal.
Making Computer Science Degrees Accessible for Women at Southern Oregon University
#ThisIsWhatAnEngineerLooksLike and #FemmeInSTEM are two popular social media hashtags populating social media platforms such as Twitter, Instagram, and Facebook to highlight women in science, technology, engineering, and mathematics fields. When one clicks either hashtag, they are met with smiling women in lab coats, behind microscopes, and exploring nature. The most underrepresented woman is the woman computer scientist. While women are underrepresented in most STEM fields, the number of women in the field of computer science has stayed far lower than any other field. According to Christianne Corbett and Catherine Hill (2015), women hold fewer than 20% of bachelor’s degrees in computer science. Representation of women declines in graduate programs, and representation declines even more in the workforce. A common issue seems to be women admitting feeling a lack of support. While the best support often starts as early as elementary school, colleges are seeking unique methods to better support women interested in computer science. By using past research and implementing new ideas, Southern Oregon University has the capability to alter their program to increase the number of women declaring computer science majors or minors, completing their degree, and succeeding once entering the workforce by acknowledging that the university must focus on strategies to both increase female enrollment and create a positive learning experience for women.
“Women support women” is not just a cute phrase one reads on an Instagram post. Female students report feeling better supported when they are surrounded by female peers. A high school in Canada started offering all-girl classes on computer science. According to Gail Crombie’s (1991) research on promoting computer science for girls in high school, girls who chose to take the all-girl class reported feeling higher levels of perceived support from their teacher and higher levels of confidence in their abilities than the girls in the mixed-gender class. Even more significant, their perceived support and confidence levels were equal to those of the boys in the mixed-gender classroom.
Eric Roberts, a professor at Reed College and formerly a professor at Stanford University, believes in making software engineering degrees as accessible as possible. Women are just as capable as men, but colleges need to recognize that not all women had access to strong computer science programs in high school like Crombie’s study demonstrated. Computing experiences before entering college is extremely split based on gender. On top of that, women have also faced years of gendered conditioning and often lost their confidence in their abilities in both mathematics and computer science (Roberts et all, 2002). Women are capable of developing the same programming skills quickly, but they need support from individuals, the computer science department, professors, and their peers.
Eric Roberts and his team at Stanford University implemented a new bridge program to support women interested in computer science; it is seeing strong results. One way that they support women is to create a pool of qualified women who can mentor future generations (Roberts et all, 2002). Stanford takes an interesting approach by including advanced undergraduate students in their pool of women mentors. Having role models at all different ages and levels of experiences allow women to seek out a mentor who best suits their current needs. While a woman with a PhD might seem like the obvious choice, pairing women with a student a year or two above a new student often gives the new student the ability to discuss issues that the older peer might have dealt with more recently such as a hostile classroom environment or online trolls or cyberbullying that female computer science students often face.
According to Roberts (2012) and his research team, common problems facing women entering computer science programs are patterns of discriminatory behavior creating a classroom environment that is unsupportive, differences in how students assessed their capabilities that correlated with their gender, scarcity of role models, and the lack of critical mass, which is the number of female students necessary to equal out the ratio of male to female students in classroom settings. Since critical mass is never reached, the computer science department often has an unwelcoming or unattractive culture to women. Roberts noted that all of these issues make recruiting and retaining women challenging.
Before the all-girl classes were offered at a Canadian high school, female students made up only 10% of the students enrolled in the class. When they began offering the all-female class, enrollment jumped significantly to 40%. This increase was consistent all three years of the study. Even if one only looks at it from an enrollment standpoint, that is a huge improvement and success (Crombie, 1991). However, Stanford decided to take a much different approach to increasing female enrollment and gauging success after considering Roberts’ opinions. Both Crombie and Roberts acknowledge that there are disproportionately small percentages of women graduating with computer science degrees. Roberts did state that low percentages were concerning, but he and his team chose to focus on and report on the actual number of women enrolled in Stanford’s computer science program (Roberts et all, 2002). Eric Roberts is well-known for supporting as many people who are interested in computer science as possible, so he believed that percentage reporting would not be as fair or as accurate as counting the actual number of enrolled students. Roberts wants the overall number of computer science majors to grow, and he was concerned that the percentage would stay disproportionately low if they only focused on that. By choosing to focus on actual numbers, Roberts and his team made room for growth while still tracing female enrollment levels.
Stanford University chose to completely rework their introduction to computer science sequence. Stanford’s introductory computer science sequence had three main goals. They wanted it to be attractive to a wide audience, they wanted to encourage every student instead of only selecting the best, and they wanted many positive role models for women. Many students taking the introductory sequence are not computer science majors. Stanford made a goal to help as many students through the sequence as possible. They believed that too many universities were focusing on making the introductory sequence too challenging and were intentionally filtering out all students except for the strongest performers (Roberts et all, 2002). Doing this benefits nobody, including the highest performing students.
Stanford University being compared to Southern Oregon University might not seem like a fair comparison since Stanford University is a research university and Southern Oregon University is a liberal arts college, but fellow liberal arts colleges have implemented similar programs with success as well. Harvey Mudd College shifted their introduction to computer science course to a Python-based experience since it is more forgiving than Java, and they also highlighted the social relevance of computer science courses. Showing practical applications attracted students who had not previously shown interest in computer science courses. Their introductory course begins by using a language created just for the class. This allows everyone to start at the same skill level while they learn together. Women show higher success rates after this course because they were able to focus on how to think like a software engineer instead of automatically spotting the students that were comfortable with the language. Harvey Mudd also introduced two different paths for students. One path was complete beginners or novices, and the other path was filled with more experienced programmers (Salim, 2012).
By offering these options, Harvey Mudd increased the percentage of women declaring computer science majors to approximately 40% when women had historically only represented 12% of Harvey Mudd’s computer science majors. This is much higher than the national average of 18%. Southern Oregon University is not Harvey Mudd, and they cannot expect to see the exact same results. However, Harvey Mudd has seen different versions of their pilot program implemented in larger universities with great success. Scalability has seen positive results at colleges like University of California, San Diego. The Harvey Mudd researchers did note that UCSD is seeing strong results and increased numbers of women declaring computer science majors (Corbett et all, 2015). Corbett and her team are not concerned about the slightly slower rate of increase since UCSD is a much larger school offering more types of programs, and Southern Oregon University could see similar results.
While Southern Oregon University does not have the ability to offer all-female classes like the Canadian high school, they could easily adapt their introductory computer science classes to be a better fit for everyone instead of only allowing the strongest to move forward with the degree. Introduction to computer science is often compared to learning a foreign language. Students are allowed to proceed to the next level of foreign language such as Spanish or Japanese with a C- or higher, and they often excel or adapt much faster in a 102 level course. Instead of creating a diverse, inclusive talent pool of computer science students, Southern Oregon University is choosing to make it more difficult on students who did not necessarily have the same experiences and support that their male peers did. Roberts listed this as one of his biggest concerns when trying to make computer science a welcoming and diverse field.
Eric Roberts (2012) has acknowledged that one of the largest issues facing colleges and universities currently is the sheer number of students interested in obtaining computer science degrees. The demand for computer science graduates is increasing at a rapid rate, and the workforce needs to be tripled. Funding for academia is an obvious concern, but Roberts believes that universities can get creative to better support students. The lack of instructors in the 1980’s caused universities to limit the number of students accepted into the computer science programs nationwide. Students are showing a renewed passion for computer science, and it does not help the students, the university, or the economy to frustrate them with limited classes or turn students away due to lack of resources.
Southern Oregon University is already following many of Roberts’s suggested best practices. One practice that Southern Oregon University could introduce is to hire undergraduate teaching assistants. Many universities are using upperclass undergraduate students to teach introductory courses, and they are having high rates of success. As more students see success in their classes, the talent pool also increases. Stanford University also noted that undergraduate teaching assistants are cheaper than graduate students, but they also have much higher success rates (Roberts et all, 2018). If Southern Oregon University chose to implement a similar program, classes could become more accessible to students. The upperclassmen would also benefit from the teaching experience, and it would add positive qualities to their resume as they enter the workforce or prepare for graduate programs.
One of the most significant issues with retention is the culture of computer science, and many women list this as their reason for changing majors. This is an issue in the workplace as well, but colleges are choosing to combat this in different ways. Harvey Mudd pays for women to attend women in computer science conferences whether they are computer science majors or not. While Harvey Mudd chooses a large national conference, they acknowledged that there are barriers to attending the Grace Hopper Celebration as well. It is not cheap, it requires extensive travel, and it sells out extremely quickly. It can also be overwhelming for students (Corbett et all, 2015). Southern Oregon University could support women in a similar fashion by either helping students attend a regional event or by hosting a regional event on campus. Regional events are often smaller and have a larger focus on students. They are also often free for students to attend. Taking even a few women can change the perception of women in computer science and give the women in attendance a better sense of what they need to succeed. This will also create a stronger program for Southern Oregon University.
Southern Oregon University is already in a position to help shift the gender gap in computer science. Being a liberal arts college means students are already being taught skills that not every engineer enters the workplace knowing such as strong writing and communication skills or the ability to understand the relationship between the project they’re working on and the social issues surrounding it. If Southern Oregon University chose to build on its many strengths, it could impact the careers and lives of many women who might not even know that computer science could be a good fit for them. Computer science is not just a technology-based program, and a liberal arts education provides an opportunity to create the most well-rounded engineer imaginable. Including women in that opportunity only increases the program’s strengths. Together, we can work together to make computer science accessible to women at Southern Oregon University.
References
Corbett, C., & Hill, C. (2015). Solving the Equation: The Variables for Women’s Success in Engineering and Computing. Washington: AAUW.
Crombie, G. (1991). Research on Young Women in Computer Science: Promoting High Technology for Girls. Women in Engineering Advisory Committee. https://files.eric.ed.gov/fulltext/ED475476.pdf
Roberts, E., Camp, T., Culler, D., Isbell, C., and Tims, J.. 2018. Rising CS Enrollments: Meeting the Challenges. In Proceedings of the 49th ACM Technical Symposium on Computer Science Education (SIGCSE ’18). Association for Computing Machinery, New York, NY, USA, 539–540. DOI:https://doi.org/10.1145/3159450.3159628
Roberts, E., Kassianidou, M., & Irani, L. (2002). Encouraging women in computer science. SIGCSE Bulletin, 34, 84–88. doi: 10.1145/543812.543837
Salim, N. (2012). The “IT” Girls: Inspiring Girls in Science, Technology, Engineering, Art, and Math [Pipelining: Attractive Programs for Women]. IEEE Women in Engineering Magazine, 6(2), 38–40. https://doi.org/10.1109/mwie.2012.2215263