ON THE NATURE AND DEVELOPMENT OF SOCIAL PRESENCE IN ONLINE COURSE DISCUSSIONS

Online learning solves a real logistical problem. It removes barriers of time and space, opens courses to students who could not otherwise attend, and requires a kind of reflective writing that face-to-face classrooms rarely demand. However, it also strips away something we often rely on without realizing it: the continuous stream of nonverbal and vocal cues that indicate other people are actually present. This loss is referred to in the research literature as the social presence problem, a term coined by Karen Swan and Li Fang Shih. Social presence, as technically defined, is the degree to which participants in computer-mediated communication feel affectively connected to one another. The concept originated with Short, Williams, and Christie, who used it to describe a property of the medium itself — how well a given channel could convey the salience of another person. Later researchers shifted the frame: social presence became partly subjective, varying from participant to participant, rather than being a fixed feature of the technology. Walther and others described a "hyperpersonalness" in online discussions, where participants project personality and build genuine feelings of connection using text alone. This raises the question: how much does that felt connection actually matter for learning? The empirical answer, gathered over about a decade before Swan and Shih's study, was consistent. Gunawardena and Zittle found that perceived social presence predicted sixty percent of the variance in students' satisfaction with a computer conference. Richardson and Swan found it predicted forty-two percent of the variability in perceived learning across seventeen online courses. Picciano found strong correlations between social presence perceptions and learning perceptions; however, there was no correlation with actual test scores. He did find that students grouped by high, medium, and low social presence differed significantly on written assignment scores, with the high group outscoring the medium group, which outscored the low group. Thus, while the connection between felt presence and outcomes was real, it was not straightforward. What those studies mostly left unexamined was a distinction that turns out to matter enormously: the difference between feeling the presence of your peers and feeling the presence of your instructor. Most prior work treated social presence as a single construct, implicitly peer-focused, reflecting the democratic ideal of online discussion as a community of equals. Richardson and Swan flagged this as a gap and recommended future work to tease the two apart. Swan and Shih set out to do exactly that. Their study encompassed four online graduate courses in educational technology: two sections of Mass Communications and Education and two sections of Educational Computing, with two instructors each teaching one section of each course. Of the ninety-one students enrolled, fifty-one completed an end-of-semester questionnaire. Both courses required weekly threaded discussions prompted by focus questions; the instructors largely maintained a restrained public presence, providing private feedback through weekly journals or email rather than frequently posting in the shared discussion space. The survey, adapted from Richardson and Swan, Gunawardena and Zittle, and Picciano, captured five perception variables on a five-point scale: perceived social presence of peers, perceived social presence of instructors, satisfaction with instructors, perceived learning from online discussions, and perceived interaction among participants. Alongside the survey, the researchers identified the five students with the highest combined social presence ratings and the five with the lowest, captured all their discussion postings, and coded them for social presence indicators such as paralanguage, emotion, humor, self-disclosure, group references, and acknowledgements. This was done using a taxonomy derived from Rourke, Anderson, Garrison, and Archer. Two coders worked independently before reaching consensus, with inter-rater reliability at ninety-four percent. Those same ten students were then interviewed by phone in sessions lasting forty-five minutes to an hour about how they prepared messages, responded to others, and experienced their instructors. Now, here's the finding that reframes the entire conversation about online course design. Perceived instructor presence correlated with instructor satisfaction at zero point eight one. Peer presence correlated with instructor satisfaction at zero point five six. When Swan and Shih ran partial correlations to isolate the unique contributions of each, the instructor presence partial remained strong at zero point seven one. The peer presence partial for instructor satisfaction dropped to negative zero point zero three — essentially zero. For perceived learning, instructor presence uniquely accounted for about twenty-four percent of the variance, while peer presence accounted for around thirteen percent once their shared influence was controlled. Together, the two predictors explained sixty-one percent of the variance in perceived learning and fifty percent of the variance in instructor satisfaction. The single area where peer presence clearly excelled was perceived interaction. Peer presence uniquely predicted how interactive students felt the discussion was, with a partial correlation of zero point four four, while the instructor presence link to interaction lost significance when peers were controlled for. Thus, the picture that emerges is specific: instructors drive satisfaction and perceived learning, while peers drive the felt sense of interaction. These are different outcomes, and conflating them in course design means optimizing for the wrong thing. Course type also mattered. Students in the Mass Communications and Education sections reported significantly lower social presence than those in Educational Computing, with a t-statistic of two point one four at forty-nine degrees of freedom and a p-value of zero point zero four. The two courses differed in discussion structure, suggesting that design choices, not just student personality or medium, shape the level of presence that develops. The qualitative findings add depth to those numbers. High-presence students used forty-two point three social presence indicators per thousand words in their posts, while low-presence students used twenty-eight point six. The gap runs across all three categories: affective indicators at twenty-six point three versus seventeen point five, interactive at ten point zero versus six point seven, and cohesive at six point zero versus four point four per thousand words. Critically, low-presence students wrote on average one point two times as many words. More volume did not translate to more presence. The interview data clarify the reasoning behind this. High-presence students were deliberate about their tone. One described her goal as making responses "sound like a normal conversation in a regular class" — choosing words carefully, adding humor, and projecting a voice into the text. They shared personal experiences because they believed it deepened learning: "You can learn a lot from people who offer to tell of their personal experiences… there has to be a way to make the online experience personable and enjoyable." These students also observed when instructors modeled a personal tone and said they mirrored it. Low-presence students took the opposite approach. They maintained a formal and "dry" tone to avoid misunderstandings, saw discussion as a task to complete rather than an exchange to engage in, and often viewed peers' social contributions as a waste of time, labeling them as "purely social commentary." Several reported that private contact with their instructor compensated for weak peer connection; one described those private exchanges as reinforcing a sense of intimacy that the public discussion never provided. Despite sometimes posting more frequently, they consistently rated perceived interaction lower than their high-presence peers. This contrast provides the clearest evidence in the study that social presence is not something that just happens to you — it's something you help construct. The specific behaviors that distinguish the two groups are learnable and teachable: self-disclosure, value statements, group references, and acknowledgements that explicitly build on what someone else has said. High-presence students engaged in these behaviors more. Low-presence students withdrew into formality and treated the discussion board like an assignment drop box. Three key takeaways emerge from this. First, instructor visibility in online discussions matters more than often assumed, and its effect on student satisfaction and perceived learning is substantial and distinct from whatever peers contribute. Second, course structure shapes social presence in measurable ways; discussion topics and grading criteria that invite personal experience appear to foster the conditions under which it develops. Third, students who feel present behave differently in ways that reinforce that presence: they write with warmth, they reference classmates by name, and they share a bit of themselves. These behaviors can be encouraged explicitly. What Swan and Shih's study ultimately shows is that text-based online discussion is not a diminished version of face-to-face conversation. It's a different communicative environment with its own logic. Connection in that environment does not happen automatically; it is built, sentence by sentence, through deliberate choices about how to present oneself on the page. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.