RESEARCH DOSSIER
Group Dynamics, Roles, and Collective Coordination
Abstract. Group dynamics is the evolving organization of interdependent behavior, roles, expectations, and cognitive resources in multi-person activity. This fundamental research monograph investigates the origins and measurement of group processes, the differentiation of task and maintenance roles, temporal coordination, task and reward interdependence, group development, shared mental models, transactive memory, information sharing, cohesion, psychological safety, conflict, composition, collective performance, and adaptive teamwork. The research integrates classic primary experiments and observational studies with theoretical frameworks and meta-analytic reviews while distinguishing association, causal interpretation, and explanatory proposals. It examines workplace, educational, distributed, and changing group contexts, and develops operational research questions concerning human–artificial coordination in the Artificial Era. The synthesis concludes that effective collective performance depends on matching role organization and information pathways to task requirements and their evolution over time. This is a critical narrative synthesis of existing scientific literature; it reports neither original experimental data nor a systematic meta-analysis.
Keywords: group dynamics; group roles; collective coordination; team cognition; task interdependence; group development; shared mental models; transactive memory; team cohesion; psychological safety; information sharing; implicit coordination; team effectiveness; role differentiation; team adaptation; distributed teams; collective intelligence; group research methods; human–artificial collaboration; Artificial Era
The Group as a Psychological System
Group dynamics describes the changing organization of interaction among people who recognize some shared membership, dependence, task, or social situation. The central scientific problem is to explain how local acts of attention, interpretation, communication, and effort produce patterned collective conduct over time. A group is therefore identified through observable relations: people respond to one another, exchange information, regulate contributions, or coordinate around consequential activities. Mere co-presence can be sufficient for certain social-influence effects but does not automatically establish a sustained task system. The study of group dynamics connects social psychology's analysis of interaction with organizational research on team design and effective collective performance. Kozlowski & Ilgen (2006) and Ilgen et al. (2005) organize this interdisciplinary territory around interdependent members, evolving processes, and context-sensitive outcomes.
Roles are recurrent expectations concerning who performs particular activities, maintains relationships, interprets uncertainty, allocates resources, or exercises influence. Collective coordination is the alignment of these differentiated activities in sequence, timing, content, and contingency. These concepts intersect because allocating responsibilities changes what each participant needs to know, while the resulting coordination experience changes what members expect of themselves and one another. Coordination may be explicit, through assignments and verbal instructions, or implicit, through anticipation derived from experience and shared task knowledge. The scientific object is the interaction of structure, behavior, cognition, and emergent group states, rather than an aggregate of isolated attitudes. Marks et al. (2001) characterizes team processes across recurring task episodes, while Rico et al. (2008) develops the theoretical importance of anticipatory coordination.
This dossier examines cohesive and loosely coupled groups undertaking a task, maintaining a shared practice, or adapting to interdependent demands. It addresses membership, role differentiation, task architecture, communication, shared cognition, group development, coordination failures, social relations, composition, measurement, and intervention. The evidentiary method is a critical narrative synthesis of identified classic empirical articles, longitudinal and field studies, theoretical analyses, and quantitative reviews. Reported associations are distinguished from experimentally manipulated mechanisms and from design recommendations. The research does not claim a systematic literature search, preregistered review, independent experiment, or new quantitative meta-analysis. The focus remains multi-person organization: individual social cognition, dyadic attunement, collective choice, and political mobilization receive attention only where they clarify the mechanisms of ongoing collective coordination.
Historical Foundations of Group Research
Scientific interest in the functioning of small groups developed through social psychology, sociometry, applied organizational investigation, and experimental analysis of interaction. Mid-twentieth-century scholarship sought a vocabulary for recurring behavior within meetings, training groups, work teams, and civic associations. Benne & Sheats (1948) distinguished task-related and group-maintenance functions, providing an early functional account of how different contributions help groups act and persist. The proposal matters historically because it treats roles as emergent patterns of participation, although the 1948 typology by itself cannot establish stable natural kinds or predict performance in every group. Historical descriptions become empirically useful when roles are connected to observable acts, assignment procedures, and outcomes that can be independently rated.
Research on group effectiveness increasingly focused on the interaction between task requirements and interpersonal process. Hackman & Morris (1975) argued that evaluating groups requires an account of the purposes interaction serves, including ways that discussion can improve or impair the use of available resources. This approach influenced later efforts to connect design characteristics, collaboration mechanisms, and effectiveness. The broader review by Mathieu et al. (2017) traces work-team research across changing organizational forms and methods. Historical claims about the Hawthorne period and laboratory traditions require distinguishing their contribution to research questions from later, stronger tests of particular mechanisms. An early observation of group behavior supplies a hypothesis and a methodological lineage; causal inference depends on the design of later studies.
Two subsequent lines substantially expanded this field. Developmental theorizing modeled the group as a temporally organized relationship rather than a fixed structure; Tuckman (1965) reviewed earlier work and proposed a sequence of common group activities. Organizational research later emphasized recurrent task cycles, adaptive coordination, and emergent shared understandings Ilgen et al. (2005). The scientific trajectory is cumulative yet revisionary. Fixed-stage accounts help organize questions about the emergence of norms and roles, while observational studies reveal trajectories that can vary with deadlines, membership transitions, and external shocks. Modern research therefore reconstructs group history from sequences of actions, observations across episodes, and changing interdependence instead of inferring a universal developmental path from group age alone.
Constructs, Levels, and Explanatory Boundaries
Group dynamics must be analyzed across at least four distinguishable units. Individuals possess expertise, motives, feelings, and role preferences. Dyads exchange signals and develop expectations of one another. The group forms a network of recurring relations, possible shared representations, and coordinated activity. An enclosing organization or institution supplies goals, resources, incentives, and formal authority. An aggregate response such as an average satisfaction score becomes a group-level construct only when theory and measurement justify that aggregation. Ilgen et al. (2005) explicitly treats teams as multilevel systems whose processes develop over time. A scientifically defensible explanation specifies whether a proposed cause operates within people, between particular partners, through group structure, or through external constraints.
Important distinctions also exist among process, emergent state, and output. A process is an activity enacted during interaction: exchanging plans, monitoring progress, confronting unexpected conditions, or assisting a teammate. An emergent state is a comparatively enduring but revisable collective property such as shared confidence, cohesion, trust, or similarity of task representations. An output is an evaluated result: correct decisions, speed, quality, member well-being, or the group's capacity to continue cooperating. Marks et al. (2001) offers a temporal taxonomy of team processes, while Kozlowski & Ilgen (2006) examines how processes and emergent states shape effectiveness. Treating all favorable responses as one broad teamwork variable conceals possible causes and consequences.
The identity of the task matters as much as the identity of the participants. A study of brainstorming, a clinical handoff, and a sports relay may involve teams of similar size but radically different information requirements and failure costs. Consequently, this monograph treats task interdependence, coordination demand, uncertainty, and time pressure as moderators rather than background decoration. Adjacent Research Dossiers can examine group decision-making and polarization, social facilitation and loafing, leadership and hierarchy, or collective action as independent focal objects. Here these phenomena enter only insofar as they reorganize participation and cooperative task execution. The distinction supports one canonical scientific object while preserving meaningful links to its neighboring fields.
Membership, Boundaries, and Social Participation
Membership establishes who is expected to contribute, who receives information, and whose actions affect collective results. Formal teams can be defined by employment records or assigned rosters, while informal groups may be constituted through repeated acknowledgment, mutual reliance, and access to shared practices. Boundary clarity has practical consequences: a missing participant at an interface may leave critical information untransmitted, whereas excessive ambiguity can create overlapping commitments. Yet permeable membership may also enable the group to recruit expertise or respond to changing demands. Mathieu et al. (2008) emphasizes that teams are embedded in organizational contexts and differ in their underlying design. The researcher must record actual interdependence and interaction rather than merely the label applied by an organization.
Participation is unevenly distributed within most groups. Some members initiate plans, others supply specialized technical knowledge, repair misunderstanding, or protect the group's social continuity. Observers can code the rate and content of turns, assignments, requests for assistance, and contributions to a shared artifact. Such observations help distinguish participation opportunities from ability or commitment. A person who speaks seldom may still provide a crucial diagnostic observation or execute a key task. Benne & Sheats (1948) offers an influential functional vocabulary; contemporary measurement should avoid equating verbal dominance with contribution value. The relevant criterion is the relationship between the behavior and the group's task and maintenance requirements.
Group boundaries also carry psychological meaning. Inclusion may affect members' willingness to disclose uncertainties and accept responsibility, while exclusion can deprive the team of otherwise available knowledge. Boundary transitions are especially revealing events for research because they change information pathways and expected role assignments. Replacement, expansion, or temporary consultation can reveal which routines depend on specific members and which are distributed across multiple participants. The ideal research design follows these changes longitudinally, recording both member-level experience and group-level outcomes. A cross-sectional judgment that a group is well integrated cannot establish how integration developed or what would happen if the composition changed. Kozlowski & Ilgen (2006) supplies the basis for examining team processes as situated and unfolding.
Task Interdependence and Group Design
A group's task can be additive, pooled, sequential, reciprocal, or intensively collaborative, and these forms place different requirements on communication and synchronization. Where members work independently and their outputs are simply added, elaborate group meetings may be unnecessary. Where one member's response determines what another can do next, timing and feedback become essential. Wageman (1995) examined the interaction of task and reward interdependence in organizational groups and reported advantages when the structure of tasks and outcomes was coherent. The study illustrates the need to treat group effectiveness as partly designed into the environment. The psychological quality of members alone cannot compensate for incompatible incentives or missing workflow interfaces.
Task architecture determines the informational burdens that roles must carry. The person responsible for final integration needs visibility into upstream contributions; specialists need a reliable way to represent uncertain diagnoses; and boundary-spanning members must translate relevant information across subtasks. Coordination can be planned through schedules, dependency maps, common procedures, and clearly defined handoffs. It can also emerge as members repeatedly adapt to breakdowns. Hackman & Wageman (2005) argues that the timing and function of coaching matter, illustrating why interventions should respond to a team's task phase rather than impose generic encouragement. A specification of task dependencies is necessary before interpreting collaboration failures as motivational deficiencies.
The productivity of interdependence is conditional. Increasing the number of participants expands potential expertise but also increases opportunities for delay, duplicated work, and communication costs. Requiring consensus or full information exchange at every step can overload a routine task; allowing total independence in a high-risk task can conceal errors until they are irreversible. This creates an empirical design problem: identify the level and pattern of collaboration that the task actually requires, then evaluate both objective results and member experience. Kozlowski & Ilgen (2006) argues that coordination should align available human resources with situational demands. Causal tests must compare feasible alternative task arrangements and account for prior skills and management conditions.
Role Differentiation and Functional Specialization
Role differentiation is the development or assignment of recurrent functional responsibilities within an interacting group. A role may specify specialized production, quality checking, facilitation, boundary communication, or care for the relationship system. Benne & Sheats (1948) presented an early taxonomy of task functions and maintenance functions, drawing attention to contributions that can be neglected when only a formal leader or final output is studied. A contemporary analysis treats these functions as hypotheses about what members actually do and why that activity matters in a particular setting. A role description becomes scientifically useful when observers can reliably identify its behavioral indicators and compare performance before and after its assignment or redistribution.
Formal assignments and enacted roles frequently diverge. A nominal coordinator may possess little current information, whereas an experienced member informally organizes the sequence of work. A designated technical expert may become a relationship mediator because colleagues trust that person's explanations. These deviations can support flexible adaptation or produce contested authority, depending on how visible and accepted they are. Mathieu et al. (2008) describes team effectiveness through reciprocal relations among inputs, mediators, and outcomes; role enactment is a practical example of a structure being continually modified by interaction. Researchers should distinguish what members are told to do, what they believe others expect, and what records of communication and action show them actually doing.
Specialization also creates vulnerabilities. An expert can become a bottleneck when only that person holds essential knowledge, a facilitator can absorb costly emotional labor, and a rigid status hierarchy can inhibit correction from peripheral members. Redundancy, cross-training, and explicit handoff criteria may reduce some of these risks, although each consumes resources. The value of stable roles must therefore be estimated against task variability, error cost, and member turnover. A study that finds superior output under assigned roles for one type of work should not be generalized to creative exploration without examining whether fixed divisions restrict the necessary information exchanges. Kozlowski & Ilgen (2006) places the effectiveness of team design within the demands of work and its changing environment.
Role Expectations, Norms, and Accountability
Role expectations are socially communicated beliefs concerning appropriate conduct and responsibility, and norms are regularities that acquire prescriptive force through group recognition and response. A group may develop expectations about preparation, speaking order, punctuality, disagreement, and help seeking before these practices are written down. Norms influence coordination by making likely actions more predictable, reducing the need to negotiate every step. They can also stabilize inefficient or exclusionary habits. The functional perspective in Benne & Sheats (1948) is a historical starting point for describing these patterns, while the dynamic account in Marks et al. (2001) suggests measuring what people do during successive episodes rather than assuming that a role label captures the whole process.
Accountability is produced by identifiable responsibility, credible feedback, and perceived consequences for contribution. In collective tasks, individual effort can become difficult to distinguish, changing incentives and the felt value of acting. The meta-analysis by Karau & Williams (1993) reviews social loafing and integrates evidence on conditions shaping motivational loss in collective performance. Its scope is important: a loss in individual effort under one pooled-output arrangement does not establish that all group work suppresses effort. Task meaning, identifiability, expected coworker contribution, and opportunities for social comparison can change the pattern. Role clarity is potentially protective when it makes consequential contributions visible without replacing collaboration with competitive monitoring.
The interaction of norms and accountability invites two complementary measurements. An observer can document actual behavior—missed handoffs, helping responses, incomplete contributions—while members report expectations and perceived fairness. Agreement among these sources is itself informative, because groups sometimes publicly endorse participation yet repeatedly reward deference. Researchers should examine temporal order: an initially weak norm might follow poor performance rather than cause it, while a new responsibility system may change both attitudes and outputs. Wageman (1995) demonstrates why reward arrangements must be analyzed with task interdependence. Longitudinal studies and interventions that alter role instructions or feedback can help separate these pathways, provided broader management changes are also recorded.
Leadership Distribution and Coordinating Authority
Leadership in group coordination includes the establishment of direction, integration of information, allocation of attention, and correction of emerging problems. Authority may be concentrated in one role, distributed among experts, or shifted according to task phase. This dossier treats leadership as a coordinating function with observable consequences, leaving the separate study of status and leader emergence to adjacent work. Salas et al. (2005) includes team leadership among the core teamwork components in a proposed integrative model, alongside performance monitoring, backup behavior, adaptability, and team orientation. The five-component structure is a theoretical synthesis whose utility depends on whether measures reliably separate its components and predict meaningful outcomes.
Distributed influence may become valuable when different people hold information needed at different moments. The coordinator of a planning session need not be the best person to direct a technical response, and a formal leader may preserve the system by making legitimate transfers of authority explicit. Coordination failures arise when several people assume another has made a decision or when expertise is ignored because it lacks formal rank. Hackman & Wageman (2005) argues that team coaching should address functions and timing; the argument supports examining whether coordination interventions are delivered when a relevant process is active. Outcomes of a leadership intervention should include information flow, timely completion, error detection, and sustained team viability rather than satisfaction alone.
Leadership distribution has costs as well as benefits. Shared influence without clear resolution procedures may multiply conflicting instructions, while strong central control may reduce monitoring by other members. The optimal arrangement depends on what members know, whether the task can be decomposed, and how quickly decisions must be made. It is therefore useful to examine authority maps as networks of requests, approvals, and escalation paths. A comparison of groups with different leadership structures should control for prior expertise and task difficulty; successful groups may simply have more capable members who find distributed leadership easier. Mathieu et al. (2017) locates such questions in the evolving organizational team research tradition, where leadership is inseparable from design and context.
Group Development and Changing Interaction
Group development concerns systematic changes in relations, expectations, tasks, and performance across time. Tuckman (1965) synthesized 50 earlier studies and proposed a sequence commonly rendered as forming, storming, norming, and performing, using different descriptions for social and task activity. The framework remains a useful historical organizer of questions about early uncertainty, conflict, cohesion, and functional differentiation. Its evidence base was a review of heterogeneous group settings rather than a prospective demonstration that every group passes through a fixed sequence. Treating the model as a probabilistic research heuristic preserves its value while allowing empirical observation to identify alternative developmental paths.
Field observation exposed nonsequential change. In an intensive study of eight naturally occurring teams, Gersick (1988) found long intervals of relative stability interrupted by major shifts in work orientation, often associated with awareness of deadlines. The study does not imply that all teams follow an identical midpoint transition; it demonstrates that temporal constraints can reorganize collective conduct and that qualitative trajectories can reveal mechanisms invisible in one-time surveys. Comparing the two traditions suggests a richer question: under what conditions does coordination become more routinized, when does the group revise its approach, and which events make a previously workable arrangement obsolete?
Membership changes, external interruptions, new task information, and accumulated successes can alter developmental trajectories. The temporal approach of Marks et al. (2001) distinguishes recurring transition processes, action processes, and interpersonal management throughout performance episodes. Development thus includes repeated re-planning and adaptation within an enduring group, not merely an initial progression toward maturity. A robust study should sample at multiple points, record task episodes and consequential events, and examine lagged relations between structure, cohesion, and performance. Groups with different task durations cannot be compared solely by elapsed calendar time; equivalent phases may be defined by decision points, deliverable cycles, or changing levels of interdependence.
Temporal Coordination and Task Episodes
Temporal coordination is the regulation of the order, pace, and synchronization of complementary activities. A team can contain all required skills and still fail because information arrives after an irreversible decision or because a handoff is omitted. Marks et al. (2001) proposes that teams shift among transition, action, and interpersonal processes, with several processes overlapping. This formulation directs researchers to record when planning occurs, when members monitor deviations, and how they recover from surprises. A time-stamped sequence of exchanges provides stronger evidence of coordination than a retrospective rating that a meeting was productive. What counts as timely depends on the task's tolerance for delay.
Transition episodes include analyzing a mission, identifying goals, choosing strategies, and specifying responsibilities. Action episodes include monitoring progress, coordinating task operations, and providing backup assistance. Interpersonal processes address conflict, emotion, motivation, and confidence across both kinds of episode. These categories organize inquiry without assuming that a task always unfolds in clean blocks. Salas et al. (2005) emphasizes closed-loop communication and mutual monitoring as mechanisms supporting cooperation; in practice the right amount of monitoring varies with danger, expertise, and workload. Recording the time cost of additional checks is necessary to assess whether they improve the net effectiveness of the collective system.
Temporal analysis should identify sequences that predict successful recovery, not merely frequencies of behaviors. For example, an unexpected constraint may trigger a request for clarification, an updated shared plan, reassignment of a role, and coordinated execution. Counting four communications as evidence of superior teamwork would overlook whether the revised plan was understood and enacted. Burke et al. (2006) theorizes team adaptation as a multilevel and multiphase process; prospective measurement can test its phases by detecting environmental perturbations and observing the subsequent actions. Such designs also need an explicit comparison: apparent improvements following a disruption may arise because the task itself became easier or because more external assistance became available.
Transactive Memory and Distributed Expertise
Transactive memory is the organization of knowledge across members together with beliefs about the location, credibility, and coordination of that knowledge. A successful team need not train every person in every field when members can recognize and consult the right expert at the right time. Lewis (2003) developed a fifteen-item measure and tested it in laboratory, MBA consulting, and technology-team samples, examining specialization, credibility, and coordination as interrelated dimensions. The contribution was methodological as well as theoretical: it turned a broad idea about shared remembering into a testable team-level construct. The scale does not independently prove causal effects or unrestricted transportability.
Distributed expertise is valuable when routes of retrieval are reliable. An expert can supply a vital answer, but that answer may remain unused if the group does not recognize who has competence, if requests are discouraged, or if information arrives after a deadline. In some situations, multiple members need redundant competence to prevent a single point of failure. DeChurch & Mesmer-Magnus (2010) shows that collective cognitive organization is linked to processes and outcomes, while Marks et al. (2001) makes clear that knowledge must be deployed during particular episodes. This suggests examining the relation between expertise maps, requests for help, response accuracy, and eventual decisions instead of treating knowledge distribution as sufficient evidence of effective coordination.
Turnover, remote communication, and accelerated technology change can disrupt expertise maps. Teams may need explicit inventories, practice rotations, and records of decisions to reestablish reliable pathways, but tools can become outdated and create false confidence. A strong design distinguishes perceived expertise from independently assessed ability and examines both correct and incorrect expert routing. It should also test whether role specialization improves output after accounting for task complexity, familiarity, and team history. The measurement program of Lewis (2003) establishes a foundation for such work, with the important qualification that evidence for a scale's internal consistency is not evidence that its score captures every relevant form of distributed cognition.
Explicit and Implicit Coordination
Explicit coordination occurs through directives, status reports, agreements, questions, and other overt communication. Implicit coordination occurs when members adjust their acts through learned expectations without a specific coordinating instruction at that moment. A skilled team may redistribute tasks as soon as one member becomes overloaded, because observed cues predict what help is needed. Rico et al. (2008) develops a team knowledge–based approach to implicit coordination and identifies shared knowledge as a theoretical enabling condition. The account provides testable propositions; conclusions about the superiority of implicit coordination require separate empirical comparisons across tasks, familiarity levels, and uncertainty.
Explicit instruction is particularly valuable when roles are new, task conditions have changed, or an unusual risk demands a common checkpoint. Implicit prediction may save time in stable environments but also propagate undetected mistaken assumptions. Therefore, the scientific question concerns adaptation between modes rather than an unconditional preference for either. Salas et al. (2005) identifies closed-loop communication as a supporting mechanism of teamwork: a message is confirmed through acknowledgment or feedback that establishes its intended meaning. Researchers can code message initiations, confirmations, requests for repair, anticipatory backup, and failures to coordinate and then relate these patterns to independent performance criteria.
A key difficulty in experimental research is separating coordination quality from the quantity of communication. A low volume of speech may reflect expert mutual anticipation or disengagement, while an elaborate conversation may indicate successful error checking or severe confusion. A task model is needed to interpret both cases. Marks et al. (2001) situates communication within episodes and interdependent functions, implying that measures should be tied to goals, deadlines, and required dependencies. Event-based observation, time-stamped digital traces, and task simulations can estimate whether the needed information reaches the person who can act on it. These analyses become stronger when they include adverse outcomes and counterfactual conditions.
Information Sharing and Collective Integration
Information sharing is the movement of relevant knowledge among group members and its integration into decisions or actions. Quantity, uniqueness, accuracy, accessibility, and timing are distinct characteristics. Mesmer-Magnus & DeChurch (2009) synthesized 72 independent studies, involving 4,795 groups and 17,279 participants, and reported a positive relation between information sharing and performance along with associations with cohesion and knowledge integration. The breadth of the meta-analysis is a substantial empirical foundation; the included studies nevertheless differ in tasks and operational definitions. The general association cannot by itself establish that requiring more verbal disclosure will improve a specific team's output.
Sharing can fail when information known to many members dominates discussion and unique facts receive limited attention. In an early experimental demonstration, Stasser & Titus (1985) gave groups different distributions of candidate information and showed how discussion and consensus requirements could favor already shared preferences over a better-informed collective judgment. That experimental paradigm is especially useful for studying the informational side of coordination, although its immediate dependent variable is group choice. The broader implication for ongoing task organization is that the allocation of speaking opportunities, requests for unique expertise, and knowledge-routing rules can shape which resources become usable.
A disciplined intervention would distinguish disclosure from incorporation. It might ask each specialist to state the evidence unique to that role, require the integrator to confirm understanding, and document whether the information changed the plan. Researchers would then compare decision accuracy, time cost, subjective burden, and later trust with a control condition. Mesmer-Magnus & DeChurch (2009) reports that task demonstrability, discussion structure, and cooperation relate to information sharing, offering variables for prospective tests. A study limited to post-session satisfaction could miss an important error: all members may feel heard while the group still fails to act on essential information. Coordination quality ultimately depends on the actionable integration of accurate knowledge.
Cohesion, Trust, and Psychological Safety
Cohesion is a family of bonds that draw members toward a group or its tasks, including interpersonal attraction and commitment to collective activity. It should be differentiated from uniformly positive mood or the mere absence of conflict. In a meta-analytic clarification, Beal et al. (2003) found that the cohesion–performance association depended on the performance criterion and workflow intensity, with stronger correlations for behavior and efficiency measures than for outcomes and effectiveness measures. The evidence also suggested separable contributions of cohesion components. A central inference is that the specific definition and measurement of both cohesion and performance strongly affect conclusions.
Psychological safety designates a shared belief that interpersonal risk taking can occur within a team. Edmondson (1999) studied 51 work teams in a manufacturing company and reported an association between psychological safety and learning behavior, with learning behavior mediating its relation to performance. This field evidence supports the relevance of a communication climate for speaking up and correcting error, although observational mediation does not establish each causal link independently. Trust, confidence in others' competence, and psychological safety should remain analytically distinct. A team may trust its technical skill but still punish a member for reporting uncertainty or disagreeing with a decision.
Cohesion can improve persistence and mutual assistance while also increasing the social costs of dissent when norms make disagreement unwelcome. The relevant empirical question is whether members can preserve commitment and cooperate while exposing information about errors, uncertainty, or alternative procedures. Gibson & Gibbs (2006) suggests that a psychologically safe communication climate can alleviate challenges associated with virtual-team conditions; the study combines qualitative and survey evidence rather than randomized assignment of climate. Consequently, intervention claims should be tested through observed voice behavior, corrective action, and objectively assessed performance. Favorable group climate is a plausible coordination resource whose value depends on its relation to the group's actual information needs.
Conflict, Disagreement, and Coordination Breakdowns
Conflict can concern interpersonal relationships, incompatible task interpretations, or disagreement about the process of allocating responsibilities. These forms are theoretically distinct but frequently correlated. De Dreu & Weingart (2003) reported negative associations of both relationship and task conflict with team performance and member satisfaction in a meta-analysis, challenging simple claims that task conflict is automatically beneficial. de Wit et al. (2012) later examined a larger and more differentiated research base, identifying conditions that can alter the pattern of associations. Together the reviews support a contextual analysis: disagreement may expose useful information, but hostile interaction, repeated ambiguity, and escalating process disputes can consume the resources required for coordinated work.
Effective conflict analysis identifies the episode and its informational consequence. A dispute over priorities may reveal incompatible assumptions, and a structured review can produce a superior plan; the same dispute can stall execution if no legitimate resolution process exists. Researchers should record who initiates disagreement, whether evidence is evaluated, how decisions are reached, and whether members continue to cooperate afterward. Marks et al. (2001) places conflict management among interpersonal processes that overlap with action and transition phases. This framework explains why measuring a group's average conflict once is a blunt instrument: the timing of disagreement relative to a deadline and the availability of repair procedures may determine its effect.
Interventions directed at conflict should target observable processes rather than rely exclusively on exhortations to be agreeable. Examples include clarification of decision rights, separation of evidence gathering from commitment, explicit protocols for challenging a plan, and post-event reviews of mistaken assumptions. Their effects cannot be guaranteed from conceptual plausibility. A randomized or staggered evaluation could compare changes in task completion, communication patterns, satisfaction, and unreported errors. Edmondson (1999) provides evidence that the perceived interpersonal climate relates to learning behavior, but transport from manufacturing teams to emergency groups or volunteer networks needs direct testing. Coordination breakdowns are multi-causal events whose diagnosis should integrate incentives, role clarity, workload, and communication.
Group Composition, Diversity, and Complementarity
Group composition describes the distribution of relevant abilities, experiences, dispositions, identities, and resources among members. An average score can conceal a critical minimum competence, an irreplaceable specialty, or a mismatch between members' strengths and task requirements. Bell (2007) meta-analyzed deep-level composition and found that associations with team performance varied by setting and operationalization. Such variation is theoretically informative: in a task requiring mutual correction, a weak point may matter more than an average; in a knowledge-integration task, diversity of expertise may expand the informational pool. A composition variable must therefore be evaluated in relation to the actual process by which members combine their resources.
Demographic diversity and functional diversity also have different causal pathways. Some differences may correlate with access to distinct experience or information, while others interact with stereotypes, institutional inequality, or subgroup formation. The existence of demographic variety does not establish that the group has exchanged or used all available perspectives. Mesmer-Magnus & DeChurch (2009) found that member heterogeneity and information distribution could affect information sharing under specific conditions; the result directs attention to interaction design as a mediator of potential informational value. Researchers should specify the diversity dimension, its measurement, and the task outcome before interpreting a pooled estimate as a general judgment about the worth of diversity.
Composition studies face substantial selection problems. Organizations may assign their strongest members to difficult projects; successful groups may retain individuals whose characteristics then appear to predict success; and high-performing teams may attract more experienced replacements. Observational correlations cannot fully separate these processes. Where assignment is feasible, randomized composition designs, controlled skill-mix simulations, and repeated performance assessment improve causal interpretation. In real organizations, multilevel models and explicit records of member selection can reduce some biases. Mathieu et al. (2008) emphasizes variation across team types, supporting comparison within meaningful task classes. The strongest design recommendation is to make complementary capabilities usable through accessible roles, information pathways, and jointly understood goals.
Group Size, Effort, and Coordination Costs
Group size influences the number of potential contributors, but it also alters communication load, visibility of individual work, and the complexity of coordination. Increasing membership can help when additional skills or parallel labor are genuinely needed, while it can hinder tasks with a narrow decision bottleneck. In additive work, an output shortfall may reflect motivational loss or flawed synchronization; these explanations demand different interventions. Karau & Williams (1993) synthesized experimental research on social loafing, making individual effort and its contingencies a major component of group performance theory. Such findings should be interpreted through task structure rather than converted into a universal rule about the ideal number of people.
Two kinds of loss must be measured separately. Process loss includes errors or delays from poor synchronization even when everyone contributes full effort. Motivational loss includes reduced exertion when effort is less identifiable, outcomes are weakly connected to contribution, or members expect others to compensate. Both may occur simultaneously. Wageman (1995) shows that task and reward arrangements can shape effective group functioning, while Kozlowski & Ilgen (2006) discusses the alignment of resources and requirements. A valid experiment could hold total expertise constant while varying interdependence, member identifiability, and communication protocols, then measure individual work rates and collective outcomes. A decline in final output alone cannot identify the operative mechanism.
Evaluating group size also requires nonlinear thinking. A small team may achieve fast alignment but lack redundancy for illness or error; a larger team may sustain continuous operation while paying more coordination cost. An appropriate dependent variable might be throughput, quality-adjusted speed, robustness to member loss, or participant learning, rather than the number of completed items alone. The investigator should test the possibility of thresholds: adding one more member can matter little until it creates a new interface or permits parallel specialization. Mathieu et al. (2017) reviews the long progression of work-team research and reinforces the importance of contextual measures. Group size is a design parameter whose effect is mediated by task distribution and coordination architecture.
Team Adaptation under Changing Conditions
Adaptation occurs when a group detects that existing routines no longer fit its situation and modifies its collective behavior. It includes recognizing an unexpected demand, revising priorities, distributing new information, reallocating resources, and validating that a new plan works. Burke et al. (2006) offers a multilevel conceptual model of adaptive team performance and discusses phases through which such changes may occur. The model is a theoretical guide that motivates empirical tests; it does not itself demonstrate that a named sequence invariably occurs. Research should define the perturbation independently of the group's response, otherwise successful adaptation can become a circular label assigned only to high-performing teams.
Adaptation depends on both stable resources and behavioral flexibility. Shared task knowledge can make deviations noticeable, while explicit communication can repair assumptions that no longer hold. An established role system may speed response in a familiar emergency yet impede change if members hesitate to cross responsibilities. Rico et al. (2008) proposes a role for implicit coordination grounded in team knowledge, suggesting that adaptation may require a transition from anticipatory routines to explicit clarification and back again. This is a testable temporal prediction. Event-based experiments can induce carefully controlled task changes and measure response latency, error recognition, information sharing, and eventual quality relative to groups receiving stable demands.
Longitudinal field studies of adaptation must consider changing task difficulty and external support. A team might appear more capable after a disruption because it received better tools, fewer constraints, or additional personnel. Relevant designs track environmental events, prior competence, intervention timing, and comparable tasks over successive episodes. Gersick (1988) demonstrates the usefulness of attending to temporal transitions, while Marks et al. (2001) provides a language for observing phases of coordination. Practical recommendations should distinguish training members to recognize changes from training them in one fixed recovery script. The capacity to reorganize roles and information pathways is an observable research target, but its transfer to a new kind of challenge remains an empirical question.
Distributed, Virtual, and Cross-Boundary Teams
Technology-mediated collaboration introduces variations in distance, time-zone overlap, channel richness, document persistence, and access to shared task state. The category virtual team conceals this diversity. Gibson & Gibbs (2006) distinguished geographic dispersion, electronic dependence, dynamic structure, and national diversity, reporting differentiated relations with innovation and the importance of communication climate in qualitative and survey research. The study warns against treating all remote teams as a single experimental condition. A geographically dispersed group with synchronous video, shared documents, and stable membership can face a different coordination problem from a colocated team with rotating shift handoffs.
Asynchronous interaction changes the temporal structure of cooperation. A written instruction may persist for later reference while lacking immediate opportunities for clarification; automated notifications may improve visibility while increasing interruption and distraction. Transactive memory and role clarity remain relevant because participants must know who can provide an answer and when that person can respond. Lewis (2003) provides an approach to measuring perceived specialization and coordination, and Mesmer-Magnus & DeChurch (2009) supports examining which information is actually shared. A strong comparison of communication modes should measure message quality, information arrival times, decisions, repair events, and work outcomes, accounting for tasks that naturally favor either synchronous or asynchronous coordination.
Distributed coordination extends beyond one formal organization to partnerships, volunteer projects, and systems involving multiple institutional authorities. Such settings often lack a single actor who can assign tasks to everyone, making boundary rules and mutual expectations especially important. Mathieu et al. (2008) emphasizes the embedding of teams in context; Kozlowski & Ilgen (2006) discusses technology and complexity as part of the modern team environment. Practical proposals include shared state records, defined escalation channels, and explicit ownership of interfaces, but their effectiveness needs task-specific evaluation. Remote technology also preserves analyzable traces, raising ethical and methodological questions concerning privacy, consent, retention, and the validity of inferring participation from visible platform activity.
Measurement, Observation, and Operationalization
Investigating group dynamics requires a measurement system matched to the proposed construct. Role differentiation can be measured by coding recurring task functions or tracing assignment records. Coordination can be indexed by the temporal fit between interdependent acts, the accuracy of handoffs, and responsiveness to unexpected constraints. Cohesion, psychological safety, and transactive memory commonly involve ratings, but rating scales should be distinguished from direct records of action. Lewis (2003) demonstrates a field-validation program for a transactive-memory instrument, while Edmondson (1999) illustrates the use of individual reports aggregated to a team level in conjunction with learning-behavior measures. Reliability and evidence of construct validity must be established for the population and purpose at hand.
Groups constitute nonindependent statistical units. Members within the same group may share context and affect one another; repeated observations of a team share history and leadership. Treating every member response or every interaction as independent can underestimate uncertainty. Appropriate analysis may require multilevel models, cluster-aware standard errors, repeated-measures models, or comparisons of intact groups. Aggregating ratings requires a theory of sharedness and empirical evidence of agreement where agreement is essential to the construct. Ilgen et al. (2005) provides the conceptual basis for distinguishing individual, team, and contextual processes. In a small sample of teams, a large number of individual respondents cannot substitute for adequate group-level replication.
Observational designs can combine audio coding, video, digital collaboration traces, structured task logs, sociometric relations, and independent performance assessment. Each channel has a blind spot: message counts overlook meaning, self-reports may reflect common perceptions rather than actual coordination, and expert ratings can be influenced by knowledge of final success. The methodological response is triangulation with preregistered coding rules where feasible, blinded outcome assessment, interrater reliability testing, and a clear timeline of exposure and behavior. Marks et al. (2001) directs analysis to temporal episodes; Mesmer-Magnus & DeChurch (2009) shows that differing operational definitions of information sharing can affect syntheses. Measurement quality therefore precedes confidence in an explanatory theory.
Experimental, Field, and Meta-Analytic Evidence
Experimental research can manipulate role assignment, information distribution, timing constraints, task interdependence, or coordination protocols while holding selected circumstances stable. Stasser & Titus (1985) demonstrates how the distribution of shared and unique information affects discussion and collective decisions under experimentally specified conditions. Such experiments identify mechanisms in a tractable task, but transferring the result to hospitals, educational groups, or international organizations requires attention to stakes, skill, duration, and institutional authority. Field studies provide ecological detail and potentially long observation periods, although selection and changing conditions complicate causal interpretation. The strongest science connects complementary designs instead of treating one methodology as universally sufficient.
Meta-analyses provide a quantitative synthesis of observed relations across studies, but they depend on construct comparability, statistical independence, selection of eligible reports, and risks of publication bias. Beal et al. (2003) shows how distinctions among cohesion components and performance criteria affect conclusions. DeChurch & Mesmer-Magnus (2010) synthesized 231 correlations from 65 independent team-cognition studies, providing a broad basis for inference about cognitive coordination. These findings are more informative than an isolated positive study, yet pooled correlations remain dependent on the limitations of the underlying research. Analysts should report moderators, uncertainty, and heterogeneity where available rather than infer that one average applies to every group.
An evidence audit should ask which variables were manipulated, which were observed, how outcomes were rated, and whether the direction of influence could be reversed. Mesmer-Magnus & DeChurch (2009) linked information sharing to performance across many studies; that finding is compatible with both better sharing improving execution and successful teams developing better sharing practices. De Dreu & Weingart (2003) and de Wit et al. (2012) also demonstrate that conclusions change when constructs, contexts, and newer evidence are incorporated. Independent replication in different settings, well-powered designs, accessible protocols, and explicit competing hypotheses are priorities. The present narrative synthesis describes source-supported patterns and their limitations without calculating new effect sizes or claiming systematic coverage of all relevant literature.
Collective Capability and the Limits of Group-Level Scores
A collective output can exceed what isolated members could accomplish when distributed resources are effectively combined. Some laboratory research also asks whether groups display a cross-task factor of collective performance. Woolley et al. (2010) reported two studies involving 699 participants organized into groups of two to five and described a general factor associated with social sensitivity and relatively equal conversational participation. The design is historically important because it operationalized collective capability over multiple tasks rather than inferring it from one successful project. The factor's interpretation remains linked to the tasks sampled and to the statistical assumptions of the analysis; it cannot be treated as an unrestricted measure of a group's intelligence across all contexts.
Group-level performance is multidimensional. Accuracy, speed, creativity, safety, member satisfaction, learning, resilience, and the capacity to cooperate again can move in different directions. A fast team may produce lower-quality work, while deliberate coordination can increase immediate time cost and reduce later errors. Mathieu et al. (2008) stresses diverse effectiveness criteria and evolving mediator models. Therefore, a single group score needs a transparent construction rule and evidence that its components meaningfully belong together. Researchers who compare teams should also ensure that evaluation tasks are comparable and that outcome raters are independent of the coordination interventions being assessed.
Collective capability also presents an aggregation problem. The highest-performing individual does not automatically define the team's output, because allocation, recognition, communication, and task constraints determine whether individual abilities can be used. Conversely, favorable group interaction cannot produce knowledge that no member or accessible resource possesses. Bell (2007) shows that composition–performance relations are moderated by setting and operationalization; DeChurch & Mesmer-Magnus (2010) documents relations between team cognition and performance. The synthetic implication is that potential resources and their coordination are separable variables. Explanatory models should jointly measure what is available, what is actually mobilized, and which downstream criterion is affected.
Interventions, Team Training, and Design Evaluation
Interventions to improve group coordination include clearer task design, role clarification, shared planning, cross-training, feedback, structured debriefing, and explicit communication protocols. Hackman & Wageman (2005) offers a theoretical model of team coaching centered on function and timing rather than a universally effective leadership style. Salas et al. (2005) proposes monitoring, backup behavior, adaptability, and coordinated communication as components of teamwork. These theories help define candidate mechanisms, but intervention effectiveness should be established through appropriately controlled trials or persuasive longitudinal comparisons. An intervention can improve perceived collaboration while leaving error rates unchanged, so evaluation must include independently measured task outcomes.
Program evaluation begins by specifying the problem. A group may have duplicated work, insufficient expert access, ambiguous decision authority, or missed temporal handoffs. Each diagnosis implies a different treatment and a different test. Role clarification can be evaluated by errors in responsibility assignment; transactive-memory training by the accuracy of expertise routing; closed-loop communication by confirmed transmission of critical information. Lewis (2003), Rico et al. (2008), and Marks et al. (2001) offer relevant construct and process frameworks. Before introducing training, investigators should examine whether basic task or incentive redesign would be more appropriate. Improving teamwork is a joint problem of human learning and system design.
Practical evidence should include the opportunity cost and potential harm of intervention. More meetings consume time; intense monitoring may reduce autonomy; rigid checklists may become obsolete when conditions change. A useful evaluation records immediate behavior, delayed performance, adaptation to novel tasks, and participant burden. Cluster-randomized designs can help when intact groups receive different conditions, but contamination between teams and leadership effects may remain. Kozlowski & Ilgen (2006) reviews intervention leverage points within the larger science of team effectiveness. Durable recommendations depend on repeated tests across task types and populations, not on the appeal of a generic toolkit.
Developmental, Cultural, and Institutional Context
The meaning and regulation of group roles vary across classrooms, sports teams, workplaces, households, and voluntary associations. The ability to coordinate also depends on developmental experience with perspective taking, instruction, turn-taking, and collective responsibility. However, adult workplace research cannot be assumed to provide an age-invariant developmental law. Tuckman (1965) drew on different kinds of groups, and Mathieu et al. (2008) highlights variation across organizational team forms. Research involving children or adolescents requires age-appropriate tasks, valid measures of participation and understanding, and attention to the authority of teachers or caretakers. Distinguishing developmental change from experience with the task is a fundamental design requirement.
Cultural context can affect expectations about hierarchy, direct disagreement, responsibility, and public recognition. Yet country labels are broad proxies for institutional histories and heterogeneous individual practices. The comparative study in Gibson & Gibbs (2006) treats national diversity as one dimension among several characteristics of distributed teams, avoiding a single undifferentiated category of virtuality. Generalization demands measurement invariance, representative sampling where possible, and observational evidence that proposed cultural mechanisms actually influence task interaction. A result from one organization may reflect its procedures and selection rules as much as the surrounding national culture.
Institutions can stabilize coordination through formal responsibility systems, shared documentation, and reliable sanctions, but they can also restrict collaboration when procedures are disconnected from work. Policies should be interpreted as contextual interventions with group-level and member-level consequences. Wageman (1995) provides an empirical example of the interplay between task and reward structure; Edmondson (1999) shows that leadership support and a shared communication climate relate to learning behavior in a specific field setting. A comparative research program should examine how conditions of employment, authority, resources, and technological access mediate group processes. Context is part of the causal system, and external validity requires specifying what context is being transported.
Human–Artificial Coordination in the Artificial Era
In the Artificial Era, collaborative activity can involve people coordinating through and with artificial systems that schedule actions, summarize information, propose plans, or take delegated steps. The psychological research question is how these arrangements change human role expectations, allocation of attention, shared task models, and accountability. A model might perform an informational function previously carried by a human specialist, yet the surrounding human group must still determine the model's reliability, access rights, and escalation paths. Salas et al. (2005) and Marks et al. (2001) provide established teamwork constructs that can be operationalized in such settings. Their empirical results concern human teams and do not, by themselves, validate strong claims about the psychological status of an artificial participant.
Artificial coordination can be studied by manipulating the location of decision authority, the transparency of machine-generated recommendations, the rate of error correction, or the timing of machine support. Investigators could compare human-only teams with teams supplied an artificial planning or memory system on matched tasks, measuring contribution distribution, common understanding, trust calibration, and final performance. Gibson & Gibbs (2006) offers relevant dimensions of electronic dependence, and Lewis (2003) suggests questions about distributed expertise. Technical audit logs can help identify reliance and correction behavior, but reliance should not be interpreted as evidence of machine understanding or experience. Such claims require separate operational definitions and independent tests.
The philosophical horizon From Homo to Artificial introduces questions about whether durable artificial participants may eventually acquire capacities that warrant new forms of psychological description. These questions belong to a research program linking Psychology of Artificial with Psychology of Human–Artificial Relations. For this dossier, their present function is to generate clear hypotheses about interaction and coordination: how changes in an artificial partner's policy affect human planning, how persistent memory alters role assignment, and how responsibility is distributed across a mixed system. Ilgen et al. (2005) supplies a general multilevel explanatory approach. Empirical findings about human trust and team behavior must remain distinguishable from authorial philosophical interpretations of Artificial Sapiens.
Critical Scientific Discussion and Unresolved Questions
Three recurrent inferential difficulties shape the literature. First, many team variables are measured through contemporaneous self-report, inviting common-method covariance and uncertainty about causal direction. Second, groups differ in tasks, composition, incentive systems, and history, so apparently similar predictors may operate through different mechanisms. Third, both processes and group states change as a consequence of performance, violating simple static assumptions. Ilgen et al. (2005) and Mathieu et al. (2008) argue for temporally aware mediator models; DeChurch & Mesmer-Magnus (2010) documents valuable aggregate cognitive associations while leaving the causal sequencing problem open. Scientific progress depends on models whose unit, time scale, and measured variables are specified before data are interpreted.
Important competing explanations remain. Does communication improve because a group trusts its members, or does trust develop because useful communication repeatedly works? Are role assignments beneficial because they clarify expectations or because skilled teams are more likely to create sensible roles? Do cohesive groups perform better, or does visible success generate cohesion? Beal et al. (2003) illustrates how conclusions depend on the outcome definition, while Edmondson (1999) provides field evidence relevant to the learning-climate pathway without fully resolving causal direction. Future studies should combine random assignment of selected interventions, repeated measurement, external performance assessment, and careful observation of coordination events.
Additional research priorities include generalization across cultural and economic settings, teams with changing membership, mixed human–artificial coordination, and the interaction of task safety with speed or creativity. These programs should publish operational definitions, preserve negative and null findings, and use independent replication where feasible. Evidence synthesis must also document search and screening procedures when claiming systematic coverage, a standard deliberately beyond this narrative study. Kozlowski & Ilgen (2006) provides an influential integrative research agenda, while Mathieu et al. (2017) situates the field's accumulated progress and persistent complexity. The most productive questions concern identifiable mechanisms under specified task conditions and the observable results of modifying them.
Conclusion
Group dynamics is the evolving organization of interdependent activity among members who create, negotiate, and revise a collective system of expectations and action. Roles locate responsibility and expertise; coordination aligns differentiated contributions in time and content; shared knowledge, trust, norms, and cohesion affect how effectively these processes operate. Their relationships are reciprocal. The review of Marks et al. (2001), Ilgen et al. (2005), and Kozlowski & Ilgen (2006) supports a multilevel and temporal understanding in which group structure shapes interaction and interaction reconstructs structure. Explaining a group's performance consequently requires observation of task requirements, available skills, actual coordination, and outcome criteria within a specified context.
The strongest evidence encompasses experimental manipulations of information distribution, field observation of team learning, psychometric work on distributed expertise, and meta-analytic estimates relating cohesion, cognitive organization, information sharing, conflict, and teamwork processes to outcomes. Stasser & Titus (1985), Edmondson (1999), Lewis (2003), Mesmer-Magnus & DeChurch (2009), and DeChurch & Mesmer-Magnus (2010) illustrate these complementary designs. Their joint significance lies in the identification of candidate mechanisms and measurable conditions, not in one universal recipe for successful groups. Different tasks demand different role distributions and coordination intensity, while the choice of performance measure changes how effectiveness is understood.
Further progress requires causal and longitudinal analyses of how groups recognize changing conditions, redirect information, redistribute responsibility, and recover from error. Questions about digitally distributed and human–artificial groups extend this program into the Artificial Era while maintaining the distinction between established empirical evidence and new hypotheses. A rigorous psychology of collective coordination measures what members can contribute, how those contributions become accessible to others, and what consequences follow from their organization. The scientific synthesis positions group dynamics as a research domain connecting interpersonal action, collective cognition, institutional design, and the temporal architecture of cooperative activity.
References
- 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
- Bell, S. T. (2007). Deep-level composition variables as predictors of team performance: A meta-analysis. Journal of Applied Psychology, 92(3), 595–615. https://doi.org/10.1037/0021-9010.92.3.595
- Benne, K. D., & Sheats, P. (1948). Functional roles of group members. Journal of Social Issues, 4(2), 41–49. https://doi.org/10.1111/j.1540-4560.1948.tb01783.x
- Burke, C. S., Stagl, K. C., Salas, E., Pierce, L., & Kendall, D. (2006). Understanding team adaptation: A conceptual analysis and model. Journal of Applied Psychology, 91(6), 1189–1207. https://doi.org/10.1037/0021-9010.91.6.1189
- De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology, 88(4), 741–749. https://doi.org/10.1037/0021-9010.88.4.741
- de Wit, F. R. C., Greer, L. L., & Jehn, K. A. (2012). The paradox of intragroup conflict: A meta-analysis. Journal of Applied Psychology, 97(2), 360–390. https://doi.org/10.1037/a0024844
- DeChurch, L. A., & Mesmer-Magnus, J. R. (2010). The cognitive underpinnings of effective teamwork: A meta-analysis. Journal of Applied Psychology, 95(1), 32–53. https://doi.org/10.1037/a0017328
- Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999
- 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/10.5465/256496
- Gibson, C. B., & Gibbs, J. L. (2006). Unpacking the concept of virtuality: The effects of geographic dispersion, electronic dependence, dynamic structure, and national diversity on team innovation. Administrative Science Quarterly, 51(3), 451–495. https://doi.org/10.2189/asqu.51.3.451
- Hackman, J. R., & Morris, C. G. (1975). Group tasks, group interaction process, and group performance effectiveness: A review and proposed integration. Advances in Experimental Social Psychology, 8, 45–99. https://doi.org/10.1016/S0065-2601(08)60248-8
- Hackman, J. R., & Wageman, R. (2005). A theory of team coaching. Academy of Management Review, 30(2), 269–287. https://doi.org/10.5465/amr.2005.16387885
- Ilgen, D. R., Hollenbeck, J. R., Johnson, M., & Jundt, D. (2005). Teams in organizations: From input-process-output models to IMOI models. Annual Review of Psychology, 56, 517–543. https://doi.org/10.1146/annurev.psych.56.091103.070250
- 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
- Kozlowski, S. W. J., & Ilgen, D. R. (2006). Enhancing the effectiveness of work groups and teams. Psychological Science in the Public Interest, 7(3), 77–124. https://doi.org/10.1111/j.1529-1006.2006.00030.x
- Lewis, K. (2003). Measuring transactive memory systems in the field: Scale development and validation. Journal of Applied Psychology, 88(4), 587–604. https://doi.org/10.1037/0021-9010.88.4.587
- Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Academy of Management Review, 26(3), 356–376. https://doi.org/10.5465/amr.2001.4845785
- Mathieu, J. E., Hollenbeck, J. R., van Knippenberg, D., & Ilgen, D. R. (2017). A century of work teams in the Journal of Applied Psychology. Journal of Applied Psychology, 102(3), 452–467. https://doi.org/10.1037/apl0000128
- Mathieu, J., Maynard, M. T., Rapp, T., & Gilson, L. (2008). Team effectiveness 1997–2007: A review of recent advancements and a glimpse into the future. Journal of Management, 34(3), 410–476. https://doi.org/10.1177/0149206308316061
- Mesmer-Magnus, J. R., & DeChurch, L. A. (2009). Information sharing and team performance: A meta-analysis. Journal of Applied Psychology, 94(2), 535–546. https://doi.org/10.1037/a0013773
- Rico, R., Sánchez-Manzanares, M., Gil, F., & Gibson, C. (2008). Team implicit coordination processes: A team knowledge–based approach. Academy of Management Review, 33(1), 163–184. https://doi.org/10.5465/amr.2008.27751276
- Salas, E., Sims, D. E., & Burke, C. S. (2005). Is there a Big Five in teamwork? Small Group Research, 36(5), 555–599. https://doi.org/10.1177/1046496405277134
- Stasser, G., & Titus, W. (1985). Pooling of unshared information in group decision making: Biased information sampling during discussion. Journal of Personality and Social Psychology, 48(6), 1467–1478. https://doi.org/10.1037/0022-3514.48.6.1467
- Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63(6), 384–399. https://doi.org/10.1037/h0022100
- Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science Quarterly, 40(1), 145–180. https://doi.org/10.2307/2393703
- Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., & Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups. Science, 330(6004), 686–688. https://doi.org/10.1126/science.1193147