Something happens in an online graduate classroom before it shows up in the gradebook.
The course shell is clean, the rubric posted, the live session structured, the discussion prompt relevant, and the assignment sequence is sound. Yet still the class drifts. Students ask for templates instead of risking judgment. They read feedback as a threat, not an invitation. They vanish after a hard module and come back with apology instead of inquiry, waiting for the instructor to rescue the very thinking graduate education exists to build.
We call this disengagement, usually.
Sometimes it is poor time management, unclear expectations, weak participation, or resistance to feedback. Such explanations are not wrong, exactly. They are just too small and reductionist in logic. They put the problem inside individual students and leave the emotional life of the classroom untheorized and unnamed. Online graduate courses are not only instructional containers. They are groups of working adults, emerging professionals, anxious scholars, first-generation graduate students, multilingual writers, career changers, and exhausted caregivers integrated into a shared task that demands visibility, judgment, uncertainty, and performance. The screen does not remove group dynamics. It just changes the costume.
Wilfred Bion’s theory of group dynamics gives online graduate faculty something engagement scholarship alone can not: a framework not for managing behavior but for improving academic performance at the level where performance breaks down. Bion’s distinction between the work group and the basic-assumption group is imperative in this analysis. One group is organized around the task, while another is organized around an unconscious defense against anxiety. These core psychoanalytic concepts give faculty a way to read what is happening underneath the surface (Bion, 1961). Silence is not only silence. Overreliance is not only neediness. Late work is not only procrastination. Conflict is not only conflict. Each is a clue, evidence really, of how the group is managing anxiety in relation to the task in front of it.
To be precise, the point is not to turn faculty into therapists; that is a category error, and worse, an ethical one. It is instructional. Faculty can learn to read group process without diagnosing students, contain anxiety without infantilizing adults, and move a class back toward disciplined work without flattening what graduate learning actually feels like from the inside.
Online Graduate Education Has a Group Problem
Online learning scholarship has spent two decades building out engagement, social presence, community of inquiries, transactional distance, persistence, and instructional design. These components are all focused on for good reason. These constructs explain why online students need real instructor presence, peer connection, structured interaction, and a clear path to participate. Recent literature confirms this distinction. Gökoğlu et al. (2024), working from 1,281 responses, found that community of inquiry and transactional distance shape behavioral engagement in online learning environments. Shehzad and Charles (2023), using a sequential explanatory mixed-methods design, found that instructor social presence moves both engagement and attrition. Martin et al. (2024) found that students rank instructor engagement strategies as especially important, while their qualitative data showed peer, instructor, self-directed, and multimodal engagement shifting by context, not uniform, and not predictable from the syllabus alone. A 30-study meta-analysis found a moderately strong relationship between teaching presence with both perceived learning and satisfaction, correlated to course length and discipline, moderating how strong that relationship gets (Caskurlu et al., 2020).
Such literature is paramount because it substantiates how online learning is relational, not merely instructional. Students are not consuming content. They are experiencing presence, distance, responsiveness, belonging, or the absence of it. Recent qualitative and practice-based work in online graduate education has pushed this paradigm further, emphasizing learning community, peer connection, collaboration, and intellectual belonging as real features of graduate persistence, not soft add-ons to it (Bloomberg et al., 2025; Scales et al., 2024). A recent multi-phase grounded-theory case study out of the Hellenic Open University identified seven interacting parameters, instructor role, timing, and what the author calls communication and climate among them, that determine whether distance collaborative learning actually works (Niari, 2025). That framework and focus of this article share a premise, even if arrived at it differently: what happens between people in a digital classroom is not incidental to learning. It is a condition of it.
Even the best-designed online course can go emotionally flat. The community of inquiry framework tells faculty to cultivate teaching presence, social presence, and cognitive presence which are all good advice depending on how far as it goes in application. Bion asks the next question: what happens when those presences get disrupted by group anxiety? What happens when the class is present but dependent? Connected but avoidant? Polite but intellectually void? Collaborative but fused around complaint? Hopeful but waiting for rescue?
These are not marginal questions. They are the room.
Bion’s Work Group and Basic-Assumption Group
Bion’s group theory starts from a rather simple distinction: groups gather to do work, but they also gather to manage anxiety. Those two purposes are not the same thing wearing different clothes. They compete. The work group organizes academic reality, task, cooperation, frustration tolerance, and learning from experience. In a graduate course, work-group functioning looks like students reading closely, taking intellectual risks, asking better questions, tolerating feedback, revising, and connecting theory to practice. The work group can disagree, feel discomfort, name confusion, all without needing the instructor to remove the difficulty first.
The basic-assumption group is a different animal. It behaves as though the task matters less than an unconscious emotional demand underneath it. Bion (1961) named three basic assumptions: dependency, fight-flight, and pairing. None of them is a personality trait. They are temporary states a group falls into and climbs back out of, often near moments of evaluation, ambiguity, identity threat, or performance demand.
Online, basic-assumption functioning gets a boost from conditions the format itself produces; compressed calendars, asynchronous misattunement, unclear cohort norms, thin informal repair, technological fatigue, the raw exposure of written performance. Niari’s (2025) framework names some of the same terrain, timing and rhythm, instructor role, communication and climate, as structural, not incidental, to whether collaborative learning holds together. In plain language: the conditions online graduate courses reliably produce are the conditions the empirical literature already treats as consequential. Bion just explains why. Students do not simply learn online. They are seen, assessed, corrected, compared, and sometimes left alone online with feedback they do not know how to metabolize.
The online class has a nervous system.
Approach and Scope
This article is a practice-based conceptual synthesis, not an empirical study. The paper integrates Bion’s (1961) group-relations theory with the empirical, mixed-methods, and qualitative literature on social presence, community of inquiry, transactional distance, collaborative learning, and graduate learning-community formation (Bloomberg et al., 2025; Caskurlu et al., 2020; Gökoğlu et al., 2024; Martin et al., 2024; Niari, 2025; Scales et al., 2024; Shehzad & Charles, 2023). The method is thematic mapping, not statistical inference: each basic assumption gets paired with an online teaching pattern the literature has already documented: overreliance tracked to unresolved transactional distance (Gökoğlu et al., 2024), withdrawal tracked to context-dependent shifts in engagement (Martin et al., 2024), deferred agency tracked to gaps in instructor and peer presence (Shehzad & Charles, 2023), and each pairing gets linked to a faculty containment move grounded in what the teaching-presence literature indicates is effective. (Caskurlu et al., 2020). The intervention matrix in Table 1 later in this article is a conceptual synthesis. It organizes existing findings through a psychodynamic lens, but not a validated instrument, as discussed in the Limitations and Future Directions sections.
Dependency: “Tell Us Exactly What You Want”
Dependency shows up when the group starts behaving as though the instructor must supply certainty, rescue, approval, and protection before learning can even start. In online graduate courses, it looks reasonable at first, almost virtuous. Students ask for clarification. They want examples. They ask for a sample paper, then a more detailed sample, then confirmation that their reading of the assignment is “right.” They want the rubric translated into a script they can just follow.
Some of this behavior is real and should be honored. Graduate faculty should not romanticize confusion; clear expectations are an equity issue, not altruism. There is still a threshold. Most faculty know it when they cross it. It is where structure stops clarifying and starts sedating. When every ambiguity must be removed before a student can think at all, the group has quietly shifted from learning to dependence. That shift is nearly always framed as more help.
The evidence backs this pattern. Transactional distance is the psychological and communicative gap a student feels from the instructor. As a variable, it significantly predicts behavioral engagement online (Gökoğlu et al., 2024). When ambiguity runs high and that gap has not closed, repeated requests for structure stop functioning as ordinary clarification-seeking. They become an attempt to collapse the distance altogether: close the gap by handing over certainty, not by staying present as a guide.
Dependency weakens academic performance because it diminishes scholarly agency from the inside. Students become accurate followers of instructions and weak interpreters of concepts. They comply with the assignment while quietly avoiding the judgment that the assignment was built to develop. The paper gets submitted. The thinking does not mature. Left alone, this pattern tracks with something else the literature identifies: instructor social presence shapes attrition as much as engagement (Shehzad & Charles, 2023). A group whose dependency goes unnamed and uncontained does not stay dependent forever. It disengages, once certainty stops arriving on schedule.
So, what does a Bion-informed response actually sound like? Not withholding support, that is not the answer either. It provides enough structure to make thinking possible while refusing to do the thinking for the student. “Here are two defensible ways to organize this paper. Choose one and justify the choice using the rubric.” Or: “I can clarify the task. I cannot remove the graduate-level judgment that the task is asking you to practice.” The intervention hands responsibility back to the learner without walking away from the learner.
Fight-Flight: The Course Becomes the Threat
Fight-flight shows up when the group behaves as though there is a danger to confront or escape. In graduate education, the danger is rare for the instructor as a person. It is what the course requires: writing that exposes underdeveloped thinking, feedback that punctures a professional self-image built over years, multicultural content that destabilizes a comfortable innocence, legal-ethical material that strips away fantasy, research methods that reveal gaps in scholarly preparation nobody wants named out loud.
Online, fight-flight rarely looks like fighting. It looks like a sudden silence after a graded assignment. Defensive emails about the rubrics. Private peer complaint that never reaches the instructor. Camera off. Intellectualization. Missed deadlines. Discussion posts are so minimal that they technically satisfy the requirement while refusing any real contact with the material. The group fights the instructor, the assignment, the platform, the policy, the readings, and the discipline itself, or it flees into absence instead. Martin et al.'s (2024) qualitative findings back this up: engagement varies sharply by pathway, peer, instructor, self-directed, multimodal, which means that when one pathway turns threatening, students do not disengage across the board. They redirect, usually to somewhere the instructor can not see.
A classroom management response tends to swing between punitive and overaccommodating, and both moves feed the basic assumption rather than interrupt it: punishment confirms threat; rescue confirms helplessness. A Bion-informed response does something different. It contains anxiety and returns the group to the work. The instructor names the moment without shaming it, thinking something like, “The first major feedback cycle can feel exposing. That is not a sign the course has gone wrong. It is the point where graduate writing starts getting honest,” and then structures revision, peer review, rubric-based self-assessment, and targeted office hours.
The group learns, slowly, that discomfort is not emergency, feedback is not attack, revision is not humiliation. That is how academic performance improves, not by lowering the task, but by growing the group’s capacity to stay inside it.
Pairing: The Future Will Save Us
Pairing appears when the group organizes itself around a fantasy. Some future person, tool, event, template, policy, module, or authority is going to resolve the present difficulty, so nobody has to resolve it now. In online graduate education, pairing is everywhere: students wait for the final paper template before starting the literature synthesis; they wait for the “real” explanation promised in a later module; they wait for AI to make the writing manageable; they wait for the instructor to post an example that will finally remove the uncertainty; they wait for one confident peer to ask the question everyone else is avoiding. Pairing is seductive because it operates like false hope. It functions as deferred agency, the group imagining that competence arrives later, intact, skipping the awkward middle where practice happens.
The generative-AI version of this fantasy deserves its own attention because it is intensifying quickly. Research on generative AI use in graduate education finds that students under real academic pressure, such as a capstone or thesis, are more likely to reach for ChatGPT or Claude or Gemini as a substitute for their own scholarship instead of as a supportive research tool (Cummins et al., 2026). Reliance on AI to “make writing manageable” is not a technology-adoption question. It is pairing, in Bion’s sense: a nonhuman authority imagined that is absorbing the group’s anxiety about performance, so the group develops resilience to tolerate it. What helps, according to institutions that have tried building trust around this rather than issuing mandates, is transparency and participation. Training that is grounded in the discipline versus imposed from outside it (Zimmer, 2026). Applied here, that means naming what AI can legitimately do: structure a source map, check clarity, but what it can never substitute for: the graduate-level judgment the task exists to build.
The broader intervention is to convert future hope into present work. Large assignments become staged-thinking tasks: a problem statement, a source map, an application paragraph, a feedback response plan, a revision note, each due before the final product exists. The instructor asks the only question that interrupts pairing: “What are you waiting for before you begin, and what can you responsibly do right now with what you already have?” Pairing does not get interrupted by cynicism but by movement.
The Bion-Informed Online Teaching Cycle
A Bion-informed approach to online graduate teaching moves through six steps. They do not constitute a checklist as the concepts are not sequential. Bion informed instructional practice is a teaching framework that is recursive, a rhythm the instructor returns to, not a box to check once. The cycle tracks the multifaceted instructor role to recent distance-collaborative-learning research has identified: cognitive guide, managerial coordinator, social facilitator, a role that shifts across a course’s lifecycle rather than a static variable (Niari, 2025).
First: observe the group. Not surveillance. Disciplined attention. What happens after feedback goes out? Who talks to whom? Which questions keep repeating? Where does the energy just disappear? Which assignments reliably produce dependency? Which topics reliably produce flight? Observation is the cognitive-guide function, and it has to come first, because nothing after it works without it.
Second: identify the likely group state. Is this a work group, or has it shifted toward dependency, fight-flight, or pairing? Observation without this step produces data without direction. Naming the state is what turns attention into diagnosis.
Third: contain the anxiety. Containment does not mean emotional caretaking minus standards; that is, comfort, not containment, and the two get confused constantly. It means the instructor holds the frame steadily enough that students can think independently inside it. Calm tone. Clear sequence. Transparent criteria. Non-defensive authority.
Fourth: clarify the academic task. What are we here to learn, demonstrate, analyze, revise, apply? The task must be named continuously, because anxiety keeps pulling attention off it. This is the design-and-organization dimension of teaching presence, the piece of instructor behavior most consistently tied to how students structure their own expectations, and in turn, to what they perceive themselves learning (Caskurlu et al., 2020).
Fifth: structure participation. Students need protocols that make scholarly work visible: breakout roles, discussion stems, peer consultation, rubric annotation, live writing, reflective check-ins, revision memos, process-based assessment. These moves line up with the facilitation and direct-instruction dimensions of teaching presence, the ones shown to correlate with perceived learning and satisfaction (Caskurlu et al., 2020).
Sixth: link process to performance. Students need to see, explicitly, that group functioning and academic outcomes are not separate categories. The capacity to tolerate feedback, sustain inquiry, ask a precise question, and revise your own thinking is academic performance.
Practice Application: Feedback as a Group Event
In online graduate education, feedback usually gets treated as a private transaction, instructor to student, closed loop. Bion suggests otherwise: feedback is a group event. After the first major paper, the entire class can shift, and not in one direction. Some students get more engaged. Others disappear. Certain individuals fight the rubric. Some ask for reassurance that nobody can give them, while there are those who quietly decide, a verdict that they are not capable of graduate-level scholarship.
A Bion-informed instructor sees this coming and designs for it. The week after the first major feedback cycle becomes a deliberate instructional moment, not damage control. The instructor names the developmental task out loud: “This is the point in the course where feedback stops being a grade explanation and starts being material for scholarly growth.” Students complete a feedback inventory, identify one repeated pattern in their own writing, revise a single paragraph in real time, and write a short note on what changed and why. The instructor does not process feelings indefinitely, feeling becomes data, and then the group goes back to work. This is consistent with what the literature already shows: direct instruction and facilitation, the two dimensions of teaching presence most within an instructor’s control, correlate with perceived learning and satisfaction, especially at the exact moments when a course demands clarification and calibration (Caskurlu et al., 2020). That is the difference between containment and comfort, in practice and not just in theory. Comfort tries to soothe academic anxieties. Containment helps the group survive it without losing the task.
Practice Application: Discussion Boards as Group Process
Discussion boards get designed as participation artifacts, almost by default: post once, reply twice, cite the readings, submit by Sunday. Students learn the ritual fast. The problem is that ritual can imitate thought convincingly enough that nobody checks.
A Bion-informed discussion design asks a different question first: what group function is this board actually serving? Is it a work-group space, where students test ideas, connect readings, and practice the language of the discipline? Or has it quietly become a dependency ritual, students writing to please the instructor? A fight-flight space, where difficult material gets managed through vague agreement? A pairing space, where everyone is waiting for one strong post to model the acceptable answer? The question matters because behavioral engagement in discussion spaces tracks with community-of-inquiry presence and transactional distance, not with the prompt format alone (Gökoğlu et al., 2024). A well-formatted prompt can still produce a pseudo-work board if no one addresses the group state beneath it.
Faculty can intervene through the prompt design itself. Instead of “reflect on the reading,” the prompt can require three moves: name the author’s claim, apply it to a real professional problem, and identify what the claim makes harder to ignore. Peer responses can require extension, complication, respectful disagreement, something beyond agreement dressed up as engagement. The discussion board starts looking less like attendance and more like thought, happening in public.
Implications for Faculty Development
Faculty are trained to design courses, grade assignments, manage policy, and use technology. They rarely get trained to read the group, even though online graduate faculty run into group-level anxiety constantly: after grades post, during identity-based content, before capstones, during research methods, in law and ethics courses, in clinical training sequences, in any class asking students to demonstrate professional judgment in public.
So, faculty development needs a reflective group-process component, and in increasingly delivered degree programs via cyberspace this is not optional. Instructors can learn to ask: What is the group avoiding? Where is dependency being invited, maybe without my meaning to invite it? What does the class do after frustration hits? How does my own anxiety shape what I do next, am I rescuing, retaliating, overexplaining, withdrawing, tightening control, right when the group needs the opposite?
Institutions that have piloted trust-based, participatory faculty development around other emotionally loaded practice changes, adopting ethically utilized generative AI, for instance, found that frontline-driven, transparent approaches build trust faster than mandates handed down from above (Zimmer, 2026). Group-process training for online graduate faculty should follow the same model: participatory, grounded in disciplinary values, not another compliance box.
This instructional approach is rigorous because it protects the task. Faculty who can contain group anxiety hold their standards without going brittle. They support students without colluding with avoidance. They hold complexity without going vague. They teach adults like adults.
Implications for Academic Performance
Academic performance is usually measured through grades, completion, retention, and demonstrated competencies. A Bion-informed framework does not replace any of that. It explains the hidden process running underneath it.
Dependency produces compliant, underdeveloped work. Fight-flight produces work that is late, defensive, shallow. Pairing produces work that is procrastinated and never benefits from real iteration. Pseudo-work produces discussion that is polished and conceptually empty. Fragmentation reduces persistence and the peer-supported learning that makes persistence possible in the first place. These read as performance problems because they are, but they are group-process problems first, and the empirical literature agrees, in its own terms: intellectual communities in online doctoral programs meaningfully support persistence and scholarly identity (Scales et al., 2024), and peer connection and collaboration sit at the center of how online graduate students experience belonging, not at the edges of it (Bloomberg et al., 2025). A fragmented group, in Bion’s terms, is a performance risk in the field’s own terms too. The intervention is not to therapize the classroom. It is to teach with enough psychological literacy to notice, early, when the group has lost the work.
Limitations and Future Directions
This article presents a practice-based framework, not an empirical validation of Bion-informed online teaching, and I want to be honest about what those limits. Future research should examine whether faculty use of group-process reflection actually improves measurable outcomes: student engagement, assignment completion, revision quality, discussion depth, persistence, and perceived belonging. Mixed-methods designs would serve this best, because the outcomes that matter here are both behavioral and interpretive. Learning analytics can show when students disappear. Qualitative inquiry can show what the disappearance meant to them. Niari’s (2025) grounded-theory work, built from multi-phase case studies across institutions, offers one methodological template other researchers could adapt to test the intervention matrix I am proposing here.
Future work should also explore how Bion-informed instructional practice intersects with culture, power, race, language, disability, first-generation status, institutional context, because faculty absolutely can misuse group-dynamics language to pathologize students or paper over structural barriers, and I would rather name that risk than pretend it away. Anxiety is never only intrapsychic. In graduate education it is economic, racialized, gendered, linguistic, technological, institutional too. The group has a nervous system. So does the institution.
Conclusion
Online graduate education does not fail only because students disengage. Sometimes it fails because faculty get taught to manage behavior without ever being taught to understand the group underneath it. Bion gives online graduate faculty language for something most of us already sense: the class can turn dependent, defensive, avoidant, hopeful, compliant, fragmented, or alive to the work. These states matter because they shape learning before learning ever shows up as performance. A student who can tolerate feedback writes differently. A class that can survive disagreement thinks differently. A cohort that can move from anxiety back to task performs differently.
The virtual room is still a room. It has climate, pressure, silence, appetite, resistance, and possibility. Faculty who learn to read it are not doing therapy. They are doing graduate teaching, with fuller attention, finally, to what was always there.