Testing and Advancing OER Design and Pedagogy: A Case from Teaching Programming to Non-STEM Majors

Abstract

This chapter explores the use of open educational resources (OERs) for social science students with minimal computational experience. The materials were developed using an approach focusing on Rapport-building, Learner-centered (scaffolded), Equitable, and Authentic in Computational Social Science learning (RELACSS). The OERs developed offer R-based modules with real-world social science data tasks structured to build skills progressively. Research was conducted to evaluate proficiency and confidence using pre-/post-tests, completion rates, and student feedback. It also assessed the usability of platforms and cloud services like Shiny apps, Google Colab, and GitHub Education for teaching CSS to social science novices. Research findings informed the transformation of RELACSS from standalone OER modules to a learning support group grounded in OERs and OER-enabled pedagogy. Findings showed skill gains, reduced anxiety, and heightened engagement. Parallel research, including an interview study and a systematic literature review, added knowledge on creating innovative OERs for beginner-level CSS.

The ability to engage with and critically evaluate data is becoming a fundamental requirement for graduates across disciplines, including disciplines outside science, technology, engineering, and mathematics (STEM) fields. Particularly, in the context of social science education, because of the increasing availability of data and research questions requiring advanced computations, social science students are increasingly expected to acquire programming and data analysis capabilities. However, non-STEM majors often encounter challenges when starting to learn data analytics, leading to a significant gap in their educational experience. Open education resources (OERs) provide an opportunity to increase the accessibility and adaptability of learning materials to meet such needs.

This chapter introduces a case of developing and researching OERs for undergraduate social science students. While the efforts described in this chapter focus on computational social science learning among social science students, the chapter aims to contribute to the broader field of OER design and pedagogy in computational education for non-STEM students.

Understanding the Challenges: Learning Barriers for Non-STEM Students

Despite the recognized importance of statistical and computational skills, many students in social science and non-STEM majors face considerable challenges in acquiring these competencies. A significant hurdle is the prevalent negative attitudes towards statistics, often intertwined with a lack of motivation and heightened levels of anxiety surrounding data analysis (Bromage et al., 2021). Another significant barrier is the perceived lack of relevance of statistical and computational methods to the core interests of social science disciplines. Students may struggle to connect abstract statistical concepts and programming techniques to the substantive questions and theoretical frameworks that drive their fields of study (Pressimone Beckowski & Torsney, 2025). Furthermore, many non-STEM students enter undergraduate programs with limited prior mathematical knowledge and skills, which can make the learning of quantitative subjects particularly daunting (Bromage et al., 2021). This disparity in mathematical preparedness can lead to difficulties in understanding fundamental statistical and computational concepts and terminology, which are often presented in an abstract manner (Douglas, 2024).

Designing OERs with the RELACSS Approach

Recognizing these challenges, a Rapport-building, Learner-centered (scaffolded), Equitable, and Authentic Computational Social Science (RELACSS) approach was developed to guide the design of OERs for beginner data analytic education. To counteract the negative attitudes and anxiety associated with learning statistics, RELACSS emphasizes the importance of building rapport between instructors and students, as well as among peers. By creating a supportive learning environment, a sense of belonging may be cultivated and barriers may be reduced.

The Learner-centered (scaffolded) aspect of RELACSS focuses on providing incremental support to students, allowing them to progress from foundational tasks to more complex analyses. This scaffolding approach ensures that students gain mastery step by step, building their confidence and competence in computational skills. The equitable component of RELACSS emphasizes the importance of providing all students with accessible resources. By making materials openly (freely) available and adaptable, RELACSS addresses the barriers of the lack of context-specific materials for non-STEM majors.

Finally, the authentic aspect of RELACSS highlights the relevance of computational skills to real-world social science problems, helping students see the value of these skills in their academic and professional lives. To further combat anxiety, rather than starting with formulas and mathematics, materials are designed to start with authentic data examples. Then, materials continue to introduce data examples and exercises that are relevant to students’ learning contexts. This relevance not only enhances student engagement but also helps students understand how computational skills can be applied to address real-world social science questions.

Using the RELACSS approach, a series of R-based beginner OER modules delivered through interactive Shiny web applications was developed^[1] . R was chosen for its prominence in social science research and its open-source nature, aligning with the OER ethos. The Shiny apps allow students to interact with code in a user-friendly environment, seeing immediate outputs without needing to grapple with complex syntax or software installation. This interactivity minimizes the intimidation factor often associated with programming. A key feature of RELACSS is its use of the World Value Survey , an open dataset spanning up to 80 countries. This resource enables students to explore topics like social attitudes or economic disparities, tailoring their analyses to questions that resonate with their interests or cultural contexts.

During the learning process, students were first guided to read a data example from the World Value Survey, demonstrated using interactive R codes and graphics. Once they were comfortable with the learning platform and the initial idea of conducting data analysis with R codes, students were guided to develop their own research questions using the World Value Survey data. Then, students were guided to learn how the codes work through modifying the codes based on prompts. After learning basic R commands for data manipulation and analysis, students then applied the skills to a final project which required them to develop their own questions and analyze the data. For details of the theoretical background of RELACSS and the technical description, see Li (2023). In this chapter, the RELACSS approach and findings were summarized and compared against a series of new parallel studies conducted on the topic.

Evaluating Impact and Usability through Implementation

The next step in the research is to implement and evaluate a set of beginner materials developed using the RELACSS approach. To maintain objectivity and rigor, a survey approach with predetermined analytic procedure was employed. Open science practices were followed by sharing all research instruments and anonymized data on Open Science Foundation website . Ethics approval was obtained from the Institutional Review Board (IRB). Students were informed about the research and provided consent to participate.

For the survey design, target outcomes were mapped to the RELACSS approach and learners’ challenges the approach aims to overcome. Corresponding measurements were then identified. The surveys thus focused on 1. Perceptions of interactivity, authenticity, and learner-centeredness of the OERs, 2. Social aspects of the classroom (sense of belonging), 3. changes in confidence, interest, knowledge, and anxiety regarding data analysis, and to probe students’ continual interest in pursuing data analytics in social science, and 4. identification with and career interest in computational social science were measured.

The survey was administered in the form of pre-test and post-test to a group of 10 students. The pre-test was administered at the beginning of the program, and the post-test was administered at the end of the program. The survey included both quantitative, Likert-style measures and qualitative, open-ended questions. Data collected before and after the program were analyzed using descriptive statistics and visualized with plots. The qualitative data were summarized.

In addition to the survey approach, students’ final projects were collected and analyzed to assess their ability to apply the skills learned in the modules. The projects were evaluated based on a rubric that assessed their understanding of data manipulation, visualization, and interpretation.

Detailed research methods and results of the evaluation are available in Li (2023). One key finding regarding the OERs is the positive review of the interactivity of the materials. Students reported that the interactive Shiny apps were easy to use and helped provide immediate feedback to enhance their learning. Additionally, the implementation of the OERs in a series of group workshop setting was well-received, with students expressing appreciation for the connection and support, as well as the relevance of the materials to their social science studies. In terms of students’ perceptions, students reported increased confidence, interest, and knowledge in data analysis and reduced anxiety. Identity and intention to continue in the data analytics field were also increased. The overall increase in engagement, both psychosocially and academically, suggested that the RELACSS approach can be helpful in addressing the challenges faced by non-STEM novices in learning data analytics.

Expanding the Scope: Parallel Research Studies

Based on the findings from the initial RELACSS evaluation, two parallel studies were conducted to deepen the understanding of OERs in CSS education. While the two studies have not been published yet, the preliminary insights obtained from the data helped inform the design and implementation of RELACSS. One of the studies was a systematic narrative review of existing literature on pedagogies and learning approaches for statistical and computational learning in undergraduate, non-STEM major contexts. For example, one identified theme from the literature was the value of incorporating peer-learning and collaborative learning into the materials and learning process.

This further reinforced the decision to implement RELACSS as a peer-learning support group. Another example is scaffolding. Many past studies used scaffolding as a pedagogical strategy to help students build their skills progressively. This finding aligned with the RELACSS approach, which emphasizes a scaffolded learning experience. The review also highlighted the importance of interactive elements in enhancing student engagement and learning outcomes, which further supported the use of Shiny apps in RELACSS. The review also highlighted the importance of interactive elements in enhancing student engagement and learning outcomes, which further supported the use of Shiny apps in RELACSS.

Another parallel study conducted was an interview study with 13 non-STEM undergraduate students. The interviews aimed to explore students’ experiences, perceived barriers, and perceptions of learning data analytics in social science contexts. Interviews were conducted by students to lessen the pressure students may feel when being interviewed by a faculty member. Preliminary coding revealed that students often describe statistical or computational learning with terms like “stressful,” “hard,” “boring,” or “confusing.” Perceptions are often influenced by previous math experiences, especially trigonometry. Students generally believed a statistics course would involve “a lot of math,” “numbers,” “graphs,” and “charts.” Many students were hesitant to answer questions in class, preferring to remain quiet, while a few students mentioned the potential benefit of collaborative projects and group discussions for understanding statistics. These descriptive student experiences aligned with the goals of RELACSS and further helped make the RELACSS approach more relevant to students’ needs.

From OERs to a Learning Ecosystem

The research findings prompted an evolution in RELACSS’s application. The project started as a standalone OER and an experimental workshop series to implement the OERs in the learning setting. With the research findings, the project has been transformed into a learning support group integrated with peer-learning. Instead of merely learning with OERs, RELACSS has become an open learning space for students with similar interests. The research also prompted some reflections on the future iterations of the project. For example, time may be limited for students to learn the materials in a workshop setting. Students also have different starting points and learning paces. Therefore, one adjustment for the following iteration was to ask students to learn the materials on their own, thus allowing them more time to interact with the materials at their own pace. Another adjustment was to offer different OER options for students to actively choose where they want to start. While the students gathered regularly to share learning progress and ask questions as the whole cohort, students were also encouraged to group with peers similar to their learning pace and interests.

Looking ahead, there are multiple ways the project can be further expanded. For example, the project can be expanded to include more advanced materials for students who have completed the beginner modules. This would allow students to continue their learning journey and build on the skills they have acquired. Additionally, the project can be expanded to include more diverse datasets and topics, allowing students to explore a wider range of social science questions. The project can also be expanded to include more collaborative learning opportunities, such as group projects or peer mentoring programs. These expansions would not only enhance the learning experience for students, but also foster a sense of community and collaboration among learners.

References

Bromage, A., Pierce, S., Reader, T., & Compton, L. (2021). Teaching statistics to non-specialists: Challenges and strategies for success. Journal of Further and Higher Education, 46(1), 46–61. https://doi.org/10.1080/0309877x.2021.1879744

Douglas, D. (2024). Teaching statistics for the social sciences using active learning: A case study. Education Sciences, 14(11), 1163. https://doi.org/10.3390/educsci14111163

Li, M. (2023). Teaching beginner-level computational social science: Interactive open education resources with learnr and shiny apps. Frontiers in Education, 8: 1130865. https://doi.org/10.3389/feduc.2023.1130865

Pressimone Beckowski, C., & Torsney, B. M. (2025). More than numbers: The relationship between belonging and engagement in an introductory statistics course. Journal of Postsecondary Student Success, 4(2), 48–80. https://doi.org/10.33009/fsop_jpss134990

About the author

name: Manyu Li, PhD

institution: University of Louisiana at Lafayette

Manyu Li is a researcher and educator specializing in STEM education technology and evaluation, with a particular focus on open education. She also investigates the influence of psychosocial factors—such as belonging and identity—on student learning outcomes and persistence. Li has contributed to multiple international, national, and state-level initiatives aimed at expanding Open Educational Resource (OER) adoption, enhancing instructional design, and fostering collaborative knowledge creation across disciplines. Li holds a Ph.D. in Psychology from University of Pittsburgh and currently serves as an Associate Professor of Psychology at the University of Louisiana at Lafayette.