Algorithms for Her
My presentation today is entitled, “Programming Women: Platform Labor and Rhetorical Education.” In this brief talk, I share insights from an ongoing qualitative study of the contemporary landscape of algorithmic literacy education for women and other underrepresented communities in information technology. First, I’ll introduce the context for the project and my guiding research question. Next, I’ll describe my theoretical framework and methodology. Finally, I’ll share preliminary results from two case studies and gesture towards further questions for the project.
While emerging networked information technologies are commonly discussed in terms of their empowering impact on users, these technologies serve to amplify existing disparities based on race, class, and gender. A 2018 report from the US Bureau of Labor Statistics shows that women comprise only twenty-six percent of the computing workforce, and this statistic drops even more precipitously for women of color. The lack of representation points to problems of access and equity in programming, contributing to the rapid growth of alternative forms of coding education as an industry. Despite the growth of the programming industry and its influence on public discourses on literacy, the enterprise of contemporary programming education for women and underrepresented communities remains understudied (Vee). As a result, current scholarship often focuses on the rhetoric embedded within algorithms without addressing the context in which these algorithms are created.
My research project addresses this gap by analyzing how algorithmic literacy is understood, taught, and practiced in sites of programming education designed for underrepresented communities in information technology, including women of color, mothers, and gender-diverse individuals. My home disciplines of rhetoric and composition and technical communication have a long history of research on the relationship between writing, technology, and education. Within these fields, researchers have examined the rhetoricity of software (Beck), critiqued conceptions of software as a neutral, rhetoric-free tool (Chun), and complicated feminist discourses of technology (Hallenbeck). The claim that computer code and software algorithms are rhetorical texts, imbued with persuasive force and shaped by the implicit assumptions of their creators, is hardly a contested belief (Johnson; Nakamura; Noble and Tynes). Estee Beck, in her study of persuasive computer algorithms, writes, “Whether it is gender or race, ableism, class or Western values of organization and logic, the suasive appeals of persuasion . . . [shape] the encoding process of writing code.” Not only does code function rhetorically, it also offers new ways for understanding rhetoric and writing. Put simply, as Annette Vee claims, programming re-codes writing. The relationship between histories of written literacy and algorithmic literacy offers a unique opportunity for researchers to contribute to interventions into equity and information technology labor.
In order to better understand the material, social, and digital contexts where algorithms are taught and created, this project surveyed an array of sites of algorithmic literacy education designed (or marketed) for underrepresented individuals in information technology. The contemporary landscape of programming education comprises of online-only educational platforms, hybrid online and in-person meetups, structured bootcamps and workshops, as well as formal courses in traditional educational settings like universities. For each of these contexts, I chose several organizations to study. I conducted participant-observer research on in-person meetings and qualitative semi-structured interviews with several instructors and learners from each group. For the digital-only educational platforms, I enrolled in the modules as a student and used close reading practices to analyze the platform’s teaching of algorithmic literacy. Organizations that I am working with include Women Who Code, Association of Women in Computing, Code for Her, GoBridge, and SkillCrush, among others.
This research is grounded in an intersectional feminist (Crenshaw) theoretical framework—a necessary practice for studies on the relationship between identity, algorithmic literacy, and labor inequity. Kimberly Scott and Patricia Garcia found that girls of color are often left out of feminist interventions into algorithmic literacy. Responding to their critique, I turned to methodologies within intersectional feminisms that foreground the positionality of research participants and highlight the ethical imperatives of qualitative research. For this presentation today, I’m choosing to focus on two case studies: the online educational platform Skillcrush and a day-long programming workshop in Go language hosted by GoBridge Toronto.
Online coding education platforms offer structured models designed to teach algorithmic literacy, and often serve as an introduction to programming for individuals coming to programming without formal education. Such sites, many with free or low-cost memberships, highlight their role in expanding access to programming education by lowering some barriers to entry for underrepresented groups. These platforms clearly exemplify modern teaching machines. Teaching machines, defined by Bill Ferster as "a way to deliver instruction by using technology that marries content and pedagogy into a self-directed experience for a learner and which relies on minimal assistance from a live instructor," attempt to disrupt traditional classroom practice, contributing to specific literacy outcomes (17).
Founded in 2012 by Adda Birnir, Skillcrush is one such online educational platform. Skillcrush’s free online week-long coding “camp” is designed to be easily accessible. Students can complete “merit badges” through brief daily tutorials. The camp is primarily browser-based and does not require students to create a developer environment. Skillcrush’s lessons use brief, three-minute videos, short readings, and multiple-choice quizzes as their primary educational content before asking students to write a line of code. By asking students to mark each task as completed, Skillcrush gamifies their online educational program. Throughout the course, students are asked to participate in the community Slack channel, from sharing introductions to workshopping their assignments. Students on Skillcrush have a few opportunities to designate their positionality in terms of experience and motivation, but those opportunities begin and end with the introductory quiz and community discussion boards. The Skillcrush learner is assumed to be a novice without a background in computer science. This assumption is particularly apparent in the interface design and approach to literacy outcomes.
Traditional programming education approaches often focus more on the syntax rather than the problem-solving process (Malik et al.). The Skillcrush online educational platform reflects this trend, with literacy outcomes focused on acquiring concepts and syntax of basic markup languages. However, Skillcrush’s platform also stresses the importance of student dispositions of confidence. A 2019 study by Meng-Jung Tsai, Ching-Yeh Wang, and Po-Fen Hsu found that “for nonexpert programming learners, the male students had signiﬁcantly higher self-eﬃcacies than the female students in developing algorithms” (1354). Student dispositions of self-efficacy are particularly important for obtaining algorithmic literacy, and Skillcrush’s interface and content work together to demystify programming. While Skillcrush and the other online educational platforms studied here demonstrate more traditional practices for teaching algorithmic literacy, my second case study, GoBridge, reveals a slightly different approach.
GoBridge is an organization that is “dedicated to building bridges that educate underrepresented communities to teach technical skills and foster diversity in Go” (Gobridge/about-us). They offer a model of algorithmic literacy education “tailored to and provided by local communities” (Gobridge/about-us). For this project, I conducted participant-observer research at a one-day programming workshop in Toronto. From the beginning, GoBridge makes their commitment to increasing access explicit, requiring potential attendees to identify as a member of at least one underrepresented community in the workshop application. The workshop was capped at twenty-eight participants, with fourteen TAs and held at the downtown office of a large tech company in Toronto. The workshop began with an explanation of the culture of the GoBridge community, highlighting the organization’s role in shaping larger programming culture. While the workshop began with a short introduction to programming in Go, the instructor explicitly denounced “sage on the stage” educational models, instead advocating for a distributed knowledge model, asking TAs and learners alike to “call him out” throughout the day in the spirit of “diversity of thought.” Participants were seated at long tables throughout the room, with TAs interspersed throughout. As a group, we worked through various problem sets over the course of the day, “learning by doing.” Participants had a variety of backgrounds, most familiar with at least one other programming language and working as developers, with a few complete beginners. Interestingly, all of the individuals I spoke too came to programming from nontraditional paths like online education, meetups, workshops, and bootcamps.
While Skillcrush continuously referred back to the financial benefits from learning on their platform, GoBridge assumed that most (if not all) of their learners were already working in tech. Instead, the Go language was positioned as first and foremost a new tool to help learners solve problems (in addition to offering better career opportunities.) The primary literacy outcomes for the workshop were twofold: first, “enough of a foundation that [participants could] self-teach,” and a student disposition characterized by a “beginner’s mind” and dismissal of ego in order to learn the problem-solving skills needed in writing algorithms.
Vee traces the move from popular literacy to standardized literacy as a homogenizing process; she asks, “will computational literacy’s current diversity be curtailed if it is converted into a standardized computer science curriculum?” and argues, “formal schooling can make literacy more accessible to students, but can also privilege some students over others and consolidate literacy around a narrow band of practices” (219). Thus far, the community-based algorithmic literacy groups have reinforced Vee’s concerns, as evidenced by the kinds of critical literacies taught by more popular coding education groups.
Ultimately, it’s too early in this study to offer broad insights on the enterprise of contemporary programming education for underrepresented communities, but these initial case studies have generated several questions I hope to consider as I move forward with the project. I am curious to see what kinds of algorithmic literacies are practiced in the other community-based meetups, workshops, and bootcamps, especially when compared to the formal sites of education. Furthermore, what learner dispositions are most frequently cultivated by these groups as they work to make programming as a field more accessible and equitable? And finally, how and to what extent do these sites of programming education function as literacy sponsors? Undoubtedly, many more questions will arise as I continue this project. I particularly welcome any questions or feedback from you all. Thank you.