Problem-based Learning as a Tool in Addressing Gender Bias

After decades of addressing the gender bias in engineering and computer fields, there are expectations, particularly by women in these fields, that the biases would have been eradicated long before 2014. However, an Implicit Association assignment 1 addressing the Gender Gap in multiple recent semesters of a Computer Ethics class produced results which the author found both surprising and disturbing in the biases reflected, and justified, by current students. As a strategy in dealing with this, Problem Based Learning (PBL) was used as the basis of a more extensive, team-based project in the Spring 2014 iteration of the class. The three-week assignment included a preliminary assessment, a group research project, an evaluation of team members, and a follow-up assessment to determine whether the project had changed any student attitudes. The paper discusses specifics of the reasons for the PBL approach, a brief description of the characteristics of Problem Based Learning, details of the multi-part assignment, results from the Spring 2014 class, and proposed refinements for future iterations. The Problem of Under-Representation Many studies have been performed by a variety of researchers trying to understand the factors that affect the lack of representation of women in science, technology, engineering, and mathematics (STEM) fields. The complex and somewhat convoluted diagram in Figure 1, from “Women’s Underrepresentation in Science: Sociocultural and Biological Considerations” in Psychological Bulletin, illustrates the multifaceted nature of the problem. It has been hypothesized that relevant factors include items such as life choices, lack of interest, effects of hormones, gatekeeper test results, lack of support, external bias, and lack of information about STEM fields, but no consensus exists as to exactly what factors contribute to the gender gap, and what roles they play. For example, it has been suggested that women’s under-representation in STEM fields is tied to cognitive differences in higher-level mathematics, but it has not been conclusively demonstrated that significant differences exist, extent of differences in areas like spatial ability, whether any differences can be remediated, and whether differences are relevant to success. 2 The numbers involved in the gender gap in computer science are substantial and growing. The gender gap has been slowly decreasing in most STEM fields, with some, such as biological sciences, showing significant gains, leading some to believe that time will solve the problem. However, the number of women in computer science peaked in 1986, and has been significantly decreasing ever since. 3 Thus, at least in the case of computer science, some positive action must be taken if the gender gap is to be closed. The shortage of women occurs at all levels, as noted in the report “Why So Few? Women in Science, Technology, Engineering, and Mathematics” published by the American Association of University Women: In elementary, middle, and high school, girls and boys take math and science courses in roughly equal numbers, and about as many girls as boys leave high school prepared to pursue science and engineering majors in college. Yet fewer women than men pursue these majors. Among first-year college students, women are much less likely than men to P ge 26255.2 say that they intend to major in science, technology, engineering, or math.... By graduation, men outnumber women in nearly every science and engineering field, and in some, such as physics, engineering, and computer science, the difference is dramatic, with women earning only 20 percent of bachelor’s degrees. Women’s representation in science and engineering declines further at the graduate level and yet again in the transition to the workplace. 4 Given the variety of factors to which the under-representation of women in STEM has been attributed, a wide range of programs have been offered to remediate the problem. Approaches include both formal and informal education opportunities, mentoring, introduction of positive role models, hands-on explorations, industry interactions, scholarship programs, internships, and inclusion of families to affect attitudes. 4 Figure 1 Factors affecting the under-representation of women in STEM fields 2 Addressing the Problem Since the first step in dealing with any problem is awareness, computer science and computer engineering programs at the University of Tennessee at Chattanooga include awareness of the gender gap in its curriculum in the required course CPSC 3610, “Ethical and Social Issues in P ge 26255.3 Computing.” The course uses readings and discussions of classic and current ethical theories, as well as current news coverage related to computer issues. The goal is to inform, explore, and shape student attitudes toward state-of-the-art ethical issues which arise in computer professions. One of the course outcomes which is regularly assessed is awareness of complex social issues such as the digital divide and the associated gender gap in computer professions. Several strategies are have been used to cover this course material. The first was the standard readings and lectures on the gender gap in STEM fields. However, this did not lend itself well to assessment of the student awareness as an outcome, as reading and listening to lectures are not quantifiable. The next approach tried was an assignment using an Implicit Association Test to gauge student attitudes toward the gender and science. This assignment, detailed in an earlier paper, 1 asked students to read a relevant chapter of the course text, then to write a paragraph on why they think that there are so few women in engineering, the sciences, computer sciences, etc., paying special attention to whether they think there is any bias involved. Next, the students were to complete the Harvard Gender-Science Implicit Association Test (IAT). Finally, students were then to write their reactions to the Implicit Association Test results, addressing whether it was what they had expected or not. A significant number of studies since the first publication of the IAT methodology in 1998 have indicated the efficacy of Implicit Association Tests in capturing underlying attitudes, resistance to faking (e.g., participants deliberately manipulating scores), and repeatability of results. Additionally, a review of IAT studies indicates that for topics of social sensitivity (e.g., racial or gender issues), the validity of IAT measures was “relatively high” compared to the attitudes self-reported by participants. 5 The expectation was that current students in the course would believe themselves to be completely unbiased in regard to gender and scientific fields, but would discover that they had a slightly greater bias than expected. The remainder of the assignment would then allow students to reflect on their previously unrecognized biases, and give opportunity for growth of awareness of gender bias in the current profession. However, the results of this assignment were not as expected. Not only did the students in CPSC 3610 demonstrate a higher automatic association of men with science than the general population results cited by Harvard, 1 the students’ written reflections were troubling. Rather than being surprised or concerned by the possible bias demonstrated by the IAT, more than half of the students in the class regarded this association of men with science as normal and expected. Student comments, cited anonymously to protect student privacy, included, “Girls just aren’t interested in stuff like computers,” “Women’s brains can’t handle the advanced math—it’s a right brain, left brain thing,” and “Women are better at nurturing than at technical things.” Reasons given for this association were similarly disturbing: “Engineering has been, and always be [sic], a male-dominated field,” and “It’s nothing to do with societal bias—it’s how girls are raised.” This last would certainly raise the question as to whether how a society raises its children is not the ultimate expression of societal bias. Student views on prospects of women’s success in the professional world also demonstrated unconscious bias One student wrote that, “[One] reason that a woman would have a hard time getting [into] and progressing through an engineering or computer science career is that sometimes men have too much pride. There are men who would not tolerate knowing that a P ge 26255.4 woman could do a better job.” Another gave as his reason for few women in engineering as related fields as, “...women usually draws [sic] maximum benefits from their employers. If employers do not want to give a lot of benefits to an employee, they would most likely hire a male. I do not really believe there is any bias involved with this because the company just does not want to spend extra money on benefits.” The students’ comments were particularly surprising given that 40% of the faculty of the Computer Science and Engineering Department is female, giving the students ample opportunity to observe successful females in computer fields. The IAT assignment’s failure to raise the awareness of students of issues related to the gender gap made a new approach necessary. As part of a learning community investigating possible uses of Problem Based Learning in the curriculum at the University of Tennessee at Chattanooga, the author began to develop a more in-depth, team-based assignment to address the topic in a different modality. Problem Based Learning Problem Based Learning is a learner-centered educational approach, which shifts the focus of education to empowered students conducting self-directed learning. In this methodology, the “...learner is mentored and encouraged to conduct research, integrate what is learned, and apply that learning to develop a viable solution to an ill-defined problem.” 6 Problem Based Learning has been in used in medical education in the U.S. for more than thirty years,