What's in a Social Computing Course: Analyzing Computer and Information Science Syllabi

To create our dataset, we chose courses that contained “social computing” in their title that were taught in the last 6 years (since 2017). As a rapidly changing field given the emergence of new social systems and social phenomena, we decided to only include recent courses (in the last 6 years). Course titles could include other terms in addition to “social computing” since we accepted those as having a slightly more specific focus. We excluded from our analysis courses solely named “CSCW”, “Online Communities”, “Computing for Social Good”, and other related fields for reasons described in our Related Work (Section 2). We centered our analysis on the unique characteristics of social computing courses for three key reasons: i) to provide a blueprint for CS faculty creating a social computing course, ii) to provide inspiration for social computing units in other CS courses, and iii) to argue for a stronger emphasis on the social dimension of computing in HCI and CS curricula.

We obtained our final list of 25 courses through two approaches. First, we started with social computing courses that we were already familiar with. From those course websites, we then searched for the ones acknowledged as inspiration. Then, we conducted multiple rounds of querying using a search engine to find: courses called “social computing”, syllabi about “social computing”, and expanding our search for specific institutions in the U.S. and internationally. When new courses acknowledged others, we also added those to our list if they fit our criteria. We stopped searching when we could not identify any more new courses.

Our analysis was conducted in multiple rounds of thematic analysis [9]. We first familiarized ourselves with each courses and all publicly available content. Then, we extracted the list of topics taught in the course as they were listed by the course instructor. We utilized this list of topics as themes already processed by the instructor as descriptions of core units for their courses. We believe that instructors have chosen meaningful titles, and as such, our analysis method organizes these topics at a higher level. We recognize that analyzing these themes means that our analysis is limited by not accounting for sub-topics taught within units. Our analysis might misrepresent the volume of the content covered on specific topics in these courses. Furthermore, we did not have an equal number of topics for each course, some describing weekly topics vs others daily, some with sub-themes. Some courses were also shorter than 16 weeks long (e.g., quarter system). Our sample might not be exhaustive of all social computing courses taught in higher education, and might be biased towards instructors who are able or willing to share their course information publicly.

Despite the challenges in normalizing course topics, we believe that presenting the number of courses addressing a topic is worthwhile for comparison so we present these values for descriptive purposes when applicable.

We then conducted an iterative approach to find emergent themes from course topics using collaborative affinity diagramming [31]. This qualitative analysis produces a hierarchical grouping of items accepted through consensus. It is a well-established HCI approach [31] that has been used for thematically analysing CS curricula [17] and learning goals [24]. This process was conducted with all the authors using Google Spreadsheets and Miro Boards. Through repeated discussions over two weeks amongst all authors, we obtained our final set of themes. We then mapped the CS 2023 Curricula (Beta Version) to the different themes taught in social computing.

Social computing courses cover a wide variety of social systems, demonstrating that the social applications of interest for the field is broad and not just limited to the study of social media platforms or work collaboration. Through our analysis, we identified twelve different application types for social computing systems discussed within social computing courses. Based on our data analysis process, these application types were obtained from course topics (either from a list of lecture topics or a list of content topics covered in the class). This means that this list is not exhaustive, for example some other types of social systems might be discussed in social computing courses but not warrant an entire lecture. On the other hand, it is noteworthy that all of these applications types have been prominent discussion topics in social computing courses. We grouped application types that were synonymous and closely related to each other. All of the key application types (i.e., social computing systems) being taught within social computing courses that we analyzed are listed in Table 1. With the exception of remote coupling or applications used by a dyad and collaboration tools (which could be used by smaller groups of people), we observed that most of the social computing courses discuss large-scale systems which connect large groups of people.

We mapped each of these applications onto Johansen's Time/Space matrix [23, 39] according to the time and space in which the human-computer interaction takes place. Most of these applications involve interactions which occur remotely, although we observed that collaboration tools may include colocated as well as remote collaboration. Depending on the specific application, the nature of interactions within commonly studied social systems could be either synchronous or asynchronous. We observe a clear focus on systems involving large-scale and remote interactions in the social computing courses we analyzed.

Given the large-scale nature of these social computing systems, students would require foundational knowledge in Computer Science concepts like social network analysis, mechanism design, data collection, Artificial Intelligence (AI) and Machine Learning (ML) to study human-computer interactions. This is evidenced by our analysis that uncovered the foundational topics that are being taught within the social computing courses that we analyzed.

In the previous sections, we discussed the application areas and sociotechnical skills taught in social computing syllabi. A third theme that we identified was that the syllabi covered different aspects of the “life cycle” of social computing systems (SCS), including: 1) design, 2) development, 3) evolution and devolution, 4) emergent phenomena, and 5) evaluation. Note that the life cycle of a SCS is iterative. Once evaluation is complete, the life cycle does not end. Rather, evaluation feeds into subsequent rounds of design, deployment, and (d)evolution, each of which may give rise to new emergent phenomena. Further, each stage itself may be iterative in nature, e.g., low- to high-fidelity prototyping and developing a minimum viable product to a full-fledged system.

The life cycle that we identified (see Table 1) loosely maps onto Iriberri and Leroy's life cycle of online communities [21] (which itself was based on the popular information systems life cycle [4]). Our SCS life cycle stages in Table 1 differs from other software life cycles by focusing on the prerogatives of social computing researchers that must take into account complex, emergent, and dynamic sociotechnical phenomena. We provide an overview of the topics covered in the life cycle, and in Section 5, we discuss additional topics that instructors could incorporate into their syllabi.

The courses in our dataset are taught as upper-level undergraduate and graduate level courses. To understand how these courses contribute to a broader computing curricula, we mapped the CS 2023 Curricula Knowledge Areas to the topics covered by social computing courses. At the moment, the CS 2023 Curricula is in its Beta Version which means that subsequent changes may occur. In its current form, the CS 2023 Curricula contains 17 Knowledge Areas. Of these, we find that 8 are not directly related to social computing as they are not addressed in any related course topics, these include: Foundations of Programming Languages, Architecture and Organization, Operating Systems, Parallel and Distributed Computing, Software Development Fundamentals, Systems Fundamentals, Algorithms and Complexity. The Knowledge Area called Networking and Communication is not directly covered in social computing courses though its premise is the core infrastructure that enables the existence of SCS. Thus, we would recommend a courses covering how the Internet works as valuable prior background. We find that 8 others are extensively covered by social computing courses.

Since social computing courses tend to be taught in as advanced courses, they assume prior knowledge in relevant aspects of computing. Through mapping social computing courses to the CS 2023 Curricula, we found that relevant knowledge areas include: HCI; AI; Software Engineering including development processes, prototyping, social APIs; Security especially as it relates to privacy; Data Management as it relates to managing large quantities of data and considering personal information in systems design; Mathematical and Statistical Foundations for big data analysis, cleaning data, data science; Graphics and Interactive Techniques such as social network visualizations; Society, Ethics, and Professionalism for example related to peer production and digital labor consequences and concerns.

The CS 2023 Curricula identifies 6 knowledge units in HCI including 1) Understanding the User: Individual goals and interactions with others, 2) Accountability and Responsibility in Design: Sustainability, security, privacy, trust, and ethics, 3) Accessibility and Inclusive Design, 4) Evaluating the Design, 5) System Design, and 6) HCI and Society, Ethics and Professionalism (SEP). These knowledge units are mapped into two packaged courses proposed in the CS 2023 Curricula, one called ‘Introduction to HCI’ focused on user-centered design and suggested to contain a group project component for CS majors and minors, another course called ‘Data Visualization’ which is suggested as an elective course. In practice, HCI is taught as a ‘living curriculum’ without a fixed prescriptive set of requirements [12], and there exists a wide variety of Introduction to HCI courses.

Based on our findings from the core features of social computing courses, we found that some aspects of the HCI KA are important topics for social computing courses such as experimental and statistical methods for human-subjects research. A multi-course sequence composed of an ’Introduction to HCI’ followed by a ’Social Computing’ course could provide an avenue for a deeper experience in HCI education. The HCI KA contains a subtopic under ‘Understanding the User’ focused on ‘Collaboration and communication’ which considers a user in a multi‐user context, however this subtopic is not covered in the example ‘Introduction to HCI’ course. This could also be an opportunity to introduce one of the ’Application Types’ of SCS as a unit in an HCI course for exploring the social dynamics shaping users’ experiences. As our findings suggest, studying users in a multi-user context requires foundational knowledge from multiple areas that are not all covered by introductory HCI courses, such as a deeper understanding of networking technologies, sociology and social psychology, and CSCW foundations.

Our analysis showed that social computing draws on many core CS disciplines. Here, we argue SCS should itself be a core discipline of CS, and that its inclusion may help broaden participation in CS.