How Groups Become Teams: Development, Cohesion, Groupthink, and the Everyday Politics of Collective Work

Organizations group people for reasons that are mostly about the work itself rather than about sociability: a single person rarely has every skill a complex task needs, deadlines often require several hands working in parallel, and some decisions improve when more than one perspective weighs in before they are made. None of that guarantees that putting people in the same room, or the same reporting line, produces anything more than a collection of individuals who happen to share a manager.

What a real group adds, once it forms, is a set of properties that do not belong to any one member: norms about how people are expected to behave, a level of attraction or commitment to the group as a whole, patterns of who talks to whom and who defers to whom, and an informal sense of who is "in" and who is still an outsider. These properties can help a group outperform the sum of its members' individual efforts, and they can just as easily get in the way. The rest of this article follows four of the most studied of these group-level properties: how groups develop over time, what cohesion does and does not buy a team, two classic failure modes that sit on opposite sides of the cohesion coin, and the informal politics that runs through nearly every group that lasts long enough to develop coalitions and history.

The most quoted description of group development did not come from an experiment. Bruce Tuckman (1965) reviewed roughly fifty existing studies of therapy and training groups and distilled a four-stage pattern: forming, in which members orient themselves, test boundaries and avoid open conflict; storming, in which disagreement over roles, structure and influence surfaces; norming, in which the group settles on accepted ways of working together; and performing, in which the group turns its attention fully toward the task because its internal structure no longer needs constant renegotiation. Twelve years later, Tuckman and Jensen (1977) added a fifth stage, adjourning, to describe groups that have a defined end point and must disengage from the task and from one another.

The names stuck because they are memorable and because most people recognize some version of the pattern from their own experience. What gets lost when the model is repeated as a checklist is how it was built. A literature synthesis of therapy and training groups is not the same kind of evidence as a controlled study of real project teams, and even within Tuckman's own review, only about half of the studies he examined showed a distinct, separate conflict stage; a number of the groups he surveyed appeared to move from forming straight into norming without visible storming at all. That detail rarely survives into the training slide. The stages describe a plausible, common pattern of interpersonal development, not a law every group must obey in the same order or at the same pace.

Connie Gersick (1988) studied something closer to what most organizations actually assemble: eight project teams with real deadlines, observed from their first meeting to their final delivery. None of them developed the way the staircase model predicts. Instead, each team settled quickly into an initial approach and then stayed remarkably inert, repeating that approach with little visible change, until it reached roughly the temporal midpoint between its first meeting and its deadline. At that midpoint, almost every team went through a concentrated burst of change: members suddenly became more open to outside information, questioned assumptions that had gone unchallenged since the first meeting, and often revised their whole approach to the task in a short window before settling back into a new, more intense period of execution toward the deadline.

Gersick's label for this pattern, borrowed from evolutionary biology, was punctuated equilibrium: long periods of stability interrupted by brief, concentrated revolutions, rather than the steady, gradual climb through fixed stages that Tuckman's model implies. The two accounts are not simply contradictory. Tuckman describes an interpersonal and structural progression; Gersick describes how a team's task strategy tends to move in relation to a deadline, regardless of where its interpersonal development happens to be. A team can be comfortably in Tuckman's norming stage and still be due for a Gersick-style midpoint jolt, because the trigger in Gersick's data was awareness of time running out, not interpersonal readiness. The practical implication is less about naming stages and more about timing: a deliberately scheduled midpoint review, bringing in outside information before the team assumes its original approach is settled, can do the work that would otherwise wait for an accidental deadline scare.

Jon Katzenbach and Douglas Smith (1993) made a distinction that is easy to state and routinely ignored in practice. A working group shares information, coordinates individual tasks and relies mainly on each member's individual accountability; its output is essentially the sum of what its members produce separately. A team, in their narrower sense, is a small number of people with complementary skills who hold themselves mutually accountable for a shared result, and whose collective work product is meant to be more than what the same individuals would have produced working in parallel. The difference is not size, title or how often the group meets. It is whether the members actually depend on one another to produce the result, and whether they are answerable to each other for it, not only to a manager above them.

This distinction matters because managers routinely call a group a "team" as a label of enthusiasm rather than as a description of how the work is structured, and then expect team-level coordination and commitment from people who are, in Katzenbach and Smith's sense, still a working group. It also cuts the other way: not every task benefits from being restructured into a team. Work that is genuinely separable, where each person's contribution can be produced and judged independently, often performs perfectly well as a working group, and forcing constant joint decision-making onto independent tasks adds coordination cost without adding the complementary-skill benefit a team is supposed to deliver. The question worth asking before investing in team-building is not "do we feel like a team" but "does this task actually require the members to depend on each other for the result."

Group cohesion, broadly, is the degree to which members are attracted to the group, committed to its goals and willing to stay in it. It is one of the most studied properties in the small-group literature, partly because it seems like an obviously desirable thing to build and partly because the evidence turns out to be more qualified than the popular advice around team-building usually admits.

Brian Mullen and Carolyn Copper's (1994) meta-analytic integration of the cohesion-performance literature found a cohesion-performance relationship that was statistically significant but small in overall magnitude, and more importantly, found that the effect was driven mainly by one component of cohesion: members' commitment to the task itself, rather than their interpersonal attraction to one another or their pride in belonging to the group. Daniel Beal, Robin Cohen, Michael Burke and Christy McLendon's (2003) later meta-analysis refined this further, distinguishing task cohesion from social cohesion and finding that task cohesion predicted performance considerably more strongly than social cohesion did. They also identified task interdependence as a moderator: cohesion mattered more for performance when the task genuinely required members to coordinate and exchange information with one another, and mattered much less when members' contributions were largely independent of each other.

A caution belongs here that Article 20 of this series raised in a different context and that applies again. Most of the underlying studies are correlational: cohesive groups tend to perform somewhat better, but the direction of that relationship is not settled by a correlation alone. A team that is already succeeding may become more cohesive as a result of shared wins, rather than cohesion being what produced the wins; stable membership, skilled leadership or a well-designed task could independently produce both cohesion and performance without either one causing the other. The honest summary of this literature is narrower than "build team spirit and performance will follow." It is closer to: cohesion built around commitment to the task, on work that genuinely requires members to depend on each other, is associated with somewhat better performance, while cohesion built mainly around liking one another personally is a weaker and less reliable lever, and the causal arrow is not fully settled either way.

Cohesion has a well-documented dark side, and it is where Irving Janis (1972, 1982) built his account of groupthink: a mode of thinking that takes over when a group's members value staying unanimous and feeling good about the group more than they value realistically testing the available alternatives. Janis developed the concept from case analysis of major US foreign-policy decisions, most famously the 1961 Bay of Pigs invasion decision, contrasted against better-managed episodes such as the Kennedy administration's later handling of the Cuban missile crisis.

Janis described eight symptoms, which fall into three clusters. Overestimation of the group produces an illusion of invulnerability and an unquestioned belief in the group's inherent morality. Closed-mindedness produces collective rationalization of warnings and a stereotyped, dismissive view of outsiders or rivals. Pressure toward uniformity produces self-censorship of private doubts, a resulting illusion that the group's apparent unanimity is real rather than manufactured, direct social pressure on anyone who voices dissent, and self-appointed "mindguards" who screen out information that might disturb the group's shared confidence. Janis treated high group cohesion as one necessary ingredient, but not a sufficient one on its own; the symptoms became likely mainly when cohesion combined with insulation from outside opinion, a directive leader who signals a preferred outcome, a lack of established procedures for weighing alternatives methodically, and high stress with little perceived hope of a better option than the one already favored.

The theory's popularity has outpaced how firmly it has been tested. Jim Esser's (1998) review, after a quarter-century of research following Janis's original work, found that historical case studies of real fiascoes tended to support the model, but that laboratory experiments designed to isolate its mechanisms gave inconsistent results, that fewer than two dozen studies had directly tested groupthink's propositions in that period, and that researchers had not agreed on a standard way to measure the antecedent conditions Janis described, which makes comparing studies difficult. Esser's own conclusion was measured rather than triumphant: it was, in his words, still too early to say with confidence whether the theory was valid as stated, needed modification, or should eventually be discarded. That is a different claim than the one usually implied when groupthink is cited as settled science, and the more defensible position for a manager to hold is that groupthink describes a real and recognizable risk pattern supported by compelling case evidence, without treating every example of quick team consensus as proof that it has occurred.

If groupthink is what can go wrong when a group becomes too comfortable with its own agreement, social loafing is what can go wrong when individual effort becomes harder to see. The phenomenon traces back to Maximilien Ringelmann, a French agricultural engineer whose rope-pulling experiments, conducted in the 1880s and published in 1913, were largely forgotten until David Kravitz and Barbara Martin (1986) tracked down and confirmed his original results: pairs pulling a rope exerted roughly 93 percent of the force two people pulling separately would produce, trios about 85 percent, and groups of eight only about 49 percent. Ringelmann's own explanation mixed two possible causes together: people in larger groups might simply coordinate their pulling less well, timing their efforts out of sync with one another, or they might individually be trying less hard.

Bibb Latane, Kipling Williams and Stephen Harkins (1979) separated the two explanations by choosing a task where physical coordination could not be the problem: group clapping and shouting, which does not require synchronized timing the way rope pulling does. They still found that per-person output fell as group size grew, which meant a real motivational effect was present, not just a coordination problem, and they coined the term social loafing for it. Steven Karau and Kipling Williams's (1993) meta-analysis of 78 studies confirmed the effect was reliable across tasks, settings and populations, with a moderate average effect size, but also found the loafing effect shrank or disappeared under specific conditions: when individual contributions remained identifiable rather than pooled anonymously, when the task was personally meaningful to the people doing it, when the group itself was highly valued by its members, and in some cases when members expected their teammates to underperform and compensated for it. Karau and Williams organized these findings into the Collective Effort Model, which treats an individual's willingness to exert effort in a group as a function of three things: whether that effort is actually expected to affect the outcome, whether the outcome is expected to bring a personally valued reward, and how much the person cares about the outcome in the first place. For a manager assembling a team, the practical lesson is not that groups automatically multiply effort, which they do not reliably do, but that the loafing risk is largely designable away through visible individual accountability and tasks people have reason to care about.

Every group that lasts long enough to develop history also develops an informal structure alongside its formal one: who tends to back whom before a meeting even starts, who controls access to information the rest of the team needs, whose opinion other members quietly defer to regardless of title. That informal structure is what organizational behavior researchers usually mean by politics, and it is genuinely contested territory in a way most of the frameworks above are not. This article keeps that discussion narrow, to how politics shows up inside groups and teams specifically; Article 25 takes up power, conflict and negotiation as fuller theories in their own right.

Henry Mintzberg (1985) defined organizational politics precisely, as "individual or group behavior that is informal, ostensibly parochial, typically divisive, and above all, in the technical sense, illegitimate," meaning not sanctioned by formal authority, accepted ideology or certified expertise. That definition reads as unambiguously negative. Jeffrey Pfeffer's (1992) argument in Managing with Power complicates it: Pfeffer treated power and the political activity used to exercise it as close to structurally unavoidable in any sizable organization, precisely because resources are scarce, goals and the right means to reach them are often genuinely ambiguous, and formal authority alone cannot settle every contested decision in advance. On that view, the real managerial question is not whether political activity can be eliminated, since it largely cannot, but whether it stays within bounds that serve the group's actual purpose or drifts into self-serving maneuvering that serves individuals at the group's expense.

The empirical literature on how politics is experienced leans toward the harsher reading. K. Michele Kacmar and Gerald Ferris's (1991) Perceptions of Organizational Politics Scale, now one of the most widely used instruments in this area, consistently finds that employees who perceive their workplace as highly political report lower job satisfaction, lower organizational commitment and higher intentions to leave, and later research using the scale has also linked perceived politics to lower supervisor-rated performance. That consistent pattern is hard to reconcile with a purely benign view of politics, but it describes perceived, experienced politics, which is not the same thing as all political behavior; employees seem to tolerate or even value some forms of informal influence, such as networking and coalition-building that visibly serve a shared goal, while reacting negatively to politics that feels arbitrary, merit-blind or aimed at personal advantage at others' expense. Gerald Ferris and his colleagues (2007) pushed this distinction further by treating the capacity to operate politically, which they called political skill, as a legitimate managerial competency built from social astuteness, networking ability, interpersonal influence and apparent sincerity, rather than treating any political behavior as inherently corrosive.

Neither camp has the better of this argument outright. The scarcity-and-ambiguity view explains why politics appears in essentially every organization that has ever been studied, including ones with careful formal structures; the perceptions-of-politics evidence explains why employees so reliably experience it as a cost when it feels illegitimate rather than earned. A reasonable position, and the one this article takes, is that the amount of political activity a group needs is probably not zero, but the amount employees will tolerate without disengaging depends heavily on whether decision rights, resource allocation and the criteria for getting ahead feel transparent and roughly merit-based rather than captured by an informal coalition. Inside a specific team, this plays out in recognizable ways: two senior members who routinely pre-align before a meeting so that the group's apparent discussion is really a ratification of a decision already made elsewhere; information that flows freely within a subgroup and more slowly to members outside it; credit for a shared result that consistently attaches to whoever is best positioned to claim it rather than whoever actually did the work. A genuinely cohesive team is not immune to any of this. Strong shared norms can suppress the more visible forms of politics, or they can simply make a dominant coalition's influence harder for outsiders, including the team's own manager, to see.

Constructed teaching case. Padma Metrics is a fictional company, and its people, figures and events are invented for teaching purposes.

None of the frameworks above gives a manager a single lever to pull. A team's apparent harmony could reflect genuine normed agreement that Tuckman's model describes, a groupthink pattern Janis's model describes, or simply a dominant informal coalition that has settled the question before the meeting starts. A quiet member could be someone who genuinely has nothing to add, someone experiencing social loafing because their individual contribution is invisible, or someone self-censoring a disagreement they do not feel safe raising. These look similar from the outside and call for different responses, and treating one as another is itself a common and costly error: coaching a self-censoring dissenter to "speak up more" without changing what happens when they do, or waiting for apparent storming to resolve on its own when what is actually happening is an entrenched coalition, tends to make the underlying problem worse rather than better.

A further honest limitation deserves to be stated rather than papered over. Much of the research behind these frameworks, including Tuckman's original review, Janis's case studies, the Ringelmann-to-Latane line of social loafing research, and the cohesion meta-analyses, draws heavily on Western, often North American, organizational and student samples. Whether the same effect sizes, or even the same mechanisms, hold with the same strength in organizations shaped by different norms around hierarchy, collectivism or resource scarcity is a genuinely open question that this research base does not settle, and claiming otherwise would be asserting more than the evidence supports.

The five phenomena covered here, team development, cohesion, groupthink, social loafing and informal politics, do not resolve into one piece of advice. They resolve into a question a manager has to keep re-asking about a specific group at a specific moment: is the current friction, or the current absence of friction, a sign of healthy development, a sign of a cohesion risk tipping into groupthink, a sign that individual effort has become invisible, or a sign that an informal coalition has already settled what the formal meeting will only ratify. The research reviewed in this article is unusually honest about its own limits, from Tuckman's own literature review finding only partial support for a distinct conflict stage, to Esser's conclusion that groupthink is still not fully validated after decades of study, to the genuine, unresolved disagreement over whether organizational politics is best treated as an avoidable pathology or a structural feature of any resource-constrained group of people. A manager who expects a clean answer from any one of these frameworks will misread their own team more often than one who treats each framework as a partial, falsifiable lens and keeps checking which lens actually fits what is happening in the room.

Select a team you have belonged to, managed, or can study through documented public sources (a sports team, a well-documented project team, or a disguised version of a workplace team). You may disguise the organization\'s and individuals\' names.

Prepare a report of 2,000 to 2,500 words that:

1. describes the team\'s task, size, tenure and degree of interdependence, and explains whether, in Katzenbach and Smith\'s sense, it functions as a working group or a team;
2. assesses the team\'s developmental history against Tuckman and Jensen\'s stages and against Gersick\'s punctuated-equilibrium pattern, stating which account fits the evidence better and why, rather than assuming either must be correct;
3. evaluates the team\'s cohesion, distinguishing task cohesion from social cohesion, and judges whether the available evidence suggests cohesion has helped performance, hurt it through a groupthink-type risk, or shows no clear causal relationship either way;
4. identifies any observable social loafing risk, explaining which specific design features of the task (identifiability of contribution, task meaningfulness, perceived dependence on others) are present or missing;
5. analyzes the team\'s informal politics, naming the main competing explanations (scarcity/ambiguity versus perceived illegitimacy) and judging, with evidence from the team itself, which explanation fits better, or whether both apply to different parts of the dynamic; and
6. proposes two specific interventions, justified by name against the frameworks above, and states one way each intervention could plausibly fail or backfire.

Use at least five academic sources, including at least three from this article\'s reference list. Every claim about the team studied must come from something you observed, documented, or could substantiate if asked; where you are uncertain, say so rather than asserting a confident diagnosis the evidence does not support.

Assessment rubric (100 marks)

1. Beal, D. J., Cohen, R. R., Burke, M. J., & McLendon, C. L. (2003). Cohesion and performance in groups: A meta-analytic clarification of construct relations. Journal of Applied Psychology, 88(6), 989–1004. https://doi.org/10.1037/0021-9010.88.6.989
2. Esser, J. K. (1998). Alive and well after 25 years: A review of groupthink research. Organizational Behavior and Human Decision Processes, 73(2-3), 116–141. https://doi.org/10.1006/obhd.1998.2758
3. Ferris, G. R., Treadway, D. C., Perrewé, P. L., Brouer, R. L., Douglas, C., & Lux, S. (2007). Political skill in organizations. Journal of Management, 33(3), 290–320. https://doi.org/10.1177/0149206307300813
4. Gersick, C. J. G. (1988). Time and transition in work teams: Toward a new model of group development. Academy of Management Journal, 31(1), 9–41. https://doi.org/10.2307/256496
5. Janis, I. L. (1972). Victims of groupthink: A psychological study of foreign-policy decisions and fiascoes. Houghton Mifflin.
6. Kacmar, K. M., & Ferris, G. R. (1991). Perceptions of Organizational Politics Scale (POPS): Development and construct validation. Educational and Psychological Measurement, 51(1), 193–205. https://doi.org/10.1177/0013164491511019
7. Karau, S. J., & Williams, K. D. (1993). Social loafing: A meta-analytic review and theoretical integration. Journal of Personality and Social Psychology, 65(4), 681–706. https://doi.org/10.1037/0022-3514.65.4.681
8. Katzenbach, J. R., & Smith, D. K. (1993). The discipline of teams. Harvard Business Review, 71(2), 111–120.
9. Kravitz, D. A., & Martin, B. (1986). Ringelmann rediscovered: The original article. Journal of Personality and Social Psychology, 50(5), 936–941. https://doi.org/10.1037/0022-3514.50.5.936
10. Mintzberg, H. (1985). The organization as political arena. Journal of Management Studies, 22(2), 133–154. https://doi.org/10.1111/j.1467-6486.1985.tb00069.x

Other sources cited

• Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes (2nd ed.). Houghton Mifflin.
• Latané, B., Williams, K., & Harkins, S. (1979). Many hands make light the work: The causes and consequences of social loafing. Journal of Personality and Social Psychology, 37(6), 822–832. https://doi.org/10.1037/0022-3514.37.6.822
• Mullen, B., & Copper, C. (1994). The relation between group cohesiveness and performance: An integration. Psychological Bulletin, 115(2), 210–227. https://doi.org/10.1037/0033-2909.115.2.210
• Pfeffer, J. (1992). Managing with power: Politics and influence in organizations. Harvard Business School Press.
• Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63(6), 384–399. https://doi.org/10.1037/h0022100
• Tuckman, B. W., & Jensen, M. A. C. (1977). Stages of small-group development revisited. Group & Organization Studies, 2(4), 419–427. https://doi.org/10.1177/105960117700200404

Verification notes: Pfeffer (1992) and Katzenbach and Smith (1993) are described from their well-documented, widely summarized central arguments rather than quoted at length, since the books themselves were not read in full for this article. Mintzberg's (1985) definition of politics is quoted directly as it is consistently reported across independent sources. All effect sizes, symptom lists and study-design details (Tuckman's ~50-study review and its partial support for storming; Gersick's eight teams and midpoint transition; Janis's eight symptoms; Esser's 25-year review findings; Ringelmann's/Kravitz and Martin's force percentages; Karau and Williams's 78-study meta-analysis; Mullen and Copper's and Beal et al.'s cohesion findings) are reported as they appear in standard citation and research-summary sources for each original publication. The Padma Metrics case and both decision-scenario vignettes are original teaching material created for this series, not real companies or events.

Next in The Human Capital MBA series: Article 25 turns to power, conflict, negotiation and organizational citizenship behavior.