Team Dynamics Concept Card Template (Psychological Safety, Team Health)

Team Dynamics Concept Card: [Team, project or course case]

A concept card for how work teams develop, decide and perform, as taught in the leadership and organisational behaviour core of a leading MBA programme: stages of development (Tuckman 1965; Tuckman and Jensen 1977) and the midpoint transition (Gersick 1988), psychological safety (Edmondson 1999), the conditions for team effectiveness (Hackman 2002), groupthink (Janis 1972), social loafing (Latané, Williams and Harkins 1979), conflict types (Jehn 1995), decision rules and a computed team-health survey summary. It is for MBA students in learning teams, project leads and managers who want to diagnose a team with evidence rather than impressions. Source basis: the original works cited below, written up in original words. The survey items are original. Tallow Creek Medical Center, Northfield Outdoor Co. and every number are fictional and were computed in Python. Replace every [bracketed] value and every row marked Example.

Document control

Contents

• When to use it
• The idea in plain language
• Formulas and notation
• Key models in more depth
• Assumptions and limits
• How managers use it
• Worked example A (fully solved)
• Worked example B (half solved)
• Exercise C (unsolved)
• Connections to other cards
• Common mistakes and misconceptions
• Self-test questions
• My notes and next steps
• Sources and further reading
• Answer key

When to use it

One-sentence rule: use this card when a group of people who share a goal and depend on each other is performing below what its members could achieve, or when you are setting one up and want to avoid the usual process losses.

The idea in plain language

Steiner (1972) put the core of team dynamics in one line: what a team actually produces equals what it could produce given its members' resources, minus process losses (coordination problems and lower motivation in groups), and good process can add gains. Everything in this card is about reducing the losses and earning the gains.

Teams change over time. Tuckman (1965) reviewed studies of small groups and proposed four stages: forming (orientation, politeness, testing what is acceptable), storming (conflict over roles, goals and influence), norming (agreement on how to work and growing cohesion) and performing (energy goes into the task). Tuckman and Jensen (1977) added adjourning, the ending of the group. Gersick (1988), observing project teams with deadlines, found a different pattern: a team sets a framework in its first meeting, keeps it for the first half of its time, then goes through a burst of change near the midpoint before working to the deadline. The practical lesson from both is to plan a deliberate reset at the midpoint.

People must feel safe to speak. Edmondson (1999) defined team psychological safety as a shared belief that the team is safe for interpersonal risk taking: asking a question, admitting a mistake, raising a concern. In her study, safety predicted learning behaviour (seeking feedback, discussing errors), which in turn predicted performance. Nembhard and Edmondson (2006) found in hospital teams that leader inclusiveness raised psychological safety and that professional status shaped who felt safe to speak. Safety is not comfort or low standards: high safety with high accountability is the learning zone.

Teams need the right conditions. Hackman (2002) argued that leaders cannot directly make a team great, but they can create five conditions that raise the odds: a real team (clear membership, real interdependence, some stability), a compelling direction, an enabling structure (task design, size and mix, norms of conduct), a supportive organisational context (rewards, information, training and resources) and expert coaching at the right moments. Wageman, Hackman and Lehman (2005) built a diagnostic survey on these conditions.

Groups make predictable errors. Janis (1972) described groupthink: highly cohesive groups under pressure, insulated from outside views and led directively, seek agreement so strongly that they stop testing alternatives. Symptoms include an illusion of invulnerability, collective rationalisation, stereotypes of outsiders, self-censorship, an illusion of unanimity, pressure on dissenters and self-appointed "mindguards". Stasser and Titus (1985) showed that groups spend most of their discussion on information everyone already knows and neglect information held by only one member. Latané, Williams and Harkins (1979) showed that individual effort falls as group size rises (social loafing); Karau and Williams (1993) found the effect is smaller when individual contributions are identifiable, the task matters to people and the group is valued.

Conflict is not one thing. Jehn (1995) separated task conflict (disagreement about the work) from relationship conflict (personal friction); Jehn and Mannix (2001) added process conflict (who does what and how). Relationship and process conflict usually hurt. Task conflict was once thought to help, but the meta-analysis by De Dreu and Weingart (2003) found it was, on average, also negatively related to performance; whether debate helps depends on trust, psychological safety and how it is handled. The goal is open, respectful debate about ideas, kept separate from personal friction.

Formulas and notation

Communication links in a team of $n$ people (why large teams coordinate badly):

Team-health scale score for respondent $j$ on a scale of $k$ items, team mean and standard deviation:

Agreement within the team, single-item form of $r_{wg}$ (James, Demaree and Wolf 1984), using the variance of a uniform "no agreement" distribution over $A$ response options:

Majority rule with $n$ independent members who are each right with probability $p$ (Condorcet's jury theorem, 1785):

Process loss (Steiner 1972): actual productivity = potential productivity - process losses (+ process gains).

Applying the single-item $r_{wg}$ to scale scores is a simplification; James, Demaree and Wolf (1984) give a multi-item version, $r_{wg(j)}$, for proper research use. For a team pulse survey, the simple form flags disagreement well enough.

Key models in more depth

Decision rules and techniques

How a team decides shapes both the quality of the decision and people's commitment to it. Vroom and Yetton (1973) framed this as a choice among styles from "leader decides alone" through "consults" to "group decides", depending on how much decision quality and member acceptance matter and on how much information the leader holds.

Condorcet's jury theorem shows the promise and the condition of majority voting: if members are independent and each is right more often than not, a larger majority is more reliable; if each is right less often than not, a larger group makes things worse. In practice the independence condition fails when members talk first and anchor on the most senior voice, which is why techniques that collect views silently before discussion help.

Social loafing and free riding

Loafing grows with group size and when individual inputs cannot be identified. Counter-measures follow from Karau and Williams (1993): make each contribution visible (named sections, a contribution log), give each person a distinct and meaningful piece, keep teams small (five to seven for most project work), and build a team people care about. The peer feedback in mba/mba-peer-feedback is a direct tool.

Diversity and information

Diverse teams hold more unshared information, which can improve decisions only if the process surfaces it. Ask each member to state what they know that others may not before discussion, assign an information owner per topic, and record dissent. Woolley, Chabris, Pentland, Hashmi and Malone (2010) reported a collective intelligence factor in small groups linked more to equal turn-taking and social sensitivity than to the average or maximum intelligence of members.

Conflict handling

Rahim (1983) measured five styles of handling interpersonal conflict, based on concern for self and for others: integrating, obliging, dominating, avoiding and compromising. Integrating (joint problem solving) is the default for task conflict; avoiding is reasonable for trivial issues or as a short cooling-off step; dominating is sometimes needed in a safety emergency, followed by an explanation.

Assumptions and limits

How managers use it

• Team launch. Agree purpose, membership, roles, norms and the decision rule in a charter (mba/mba-learning-team-charter). Hackman's conditions are a checklist for the launch.
• Midpoint review. Schedule it at about half of the project time: rerun the survey, review the plan and change what is not working, using Gersick's finding that teams are most open to change then.
• Psychological safety moves. Frame work as learning ("we have not done this before, I need your eyes on it"), acknowledge your own fallibility, ask questions, and respond productively to bad news.
• Meeting design. Silent idea writing before discussion, a rotating devil's advocate for big choices, the most senior person speaks last, and a decision log recording the rule used.
• Conflict routine. Name the type (task, relationship, process), deal with process issues in writing (roles, deadlines), take relationship issues one to one first, and escalate according to the charter.
• Clinical and operational teams. Briefings, structured handovers (healthcare/sbar) and debriefs after events apply the same ideas under time pressure.

Worked example A (fully solved): a discharge improvement team

Situation (fictional). Tallow Creek Medical Center formed a six-person team to cut the time from "medically ready for discharge" to the patient leaving the ward: a ward nurse, a hospitalist physician, a pharmacist, a case manager, an IT analyst and a patient-access coordinator. Four weeks in, the team had missed its first milestone. The facilitator ran an anonymous survey of 15 original items, three per dimension, on a 1 (strongly disagree) to 7 (strongly agree) scale, and computed each respondent's scale score as the average of the three items. Relationship conflict is a reverse dimension: lower is better.

Survey dimensions (original items, one example each).

Raw scale scores (respondents R1 to R6).

Summary. Favourable means a score of 5 or more (3 or less for the reverse dimension). The flag rule, agreed with the team in advance: discuss a dimension if its mean is below 4.5 (above 3.5 for the reverse one) or if $r_{wg}$ is below 0.70.

Reading the result. Three lessons come out of the numbers. First, the mean hides a person: psychological safety averages 4.90, which looks acceptable, but one respondent scores 2.7 and agreement is low. Without R4 the mean would be 5.34. The same respondent reports high relationship conflict and the lowest workload fairness, so one member is having a very different experience of the team. The facilitator must not try to identify R4; the team discusses the pattern. Second, direction is unclear for everyone: clarity has the lowest mean (4.00) with high agreement (0.94), a shared problem that fits the missed milestone and needs a leader action, not a debate. Third, debate is borderline (4.50) and two scores sit at or below 4.3, which suggests some members hold back views.

Actions agreed (Example). (1) The sponsor restates the goal as "median ready-to-leave time below 2 hours on two pilot wards within 10 weeks", with weekly data. (2) The team adopts silent written input before decisions and a rotating devil's advocate. (3) Workload is rebuilt as a named task list with owners, reviewed weekly. (4) The physician lead, the most senior member, speaks last in option discussions. Re-survey in six weeks.

Decision rule check. The team had been voting by simple majority after open discussion. With five independent voters each right 65% of the time, the probability that the majority is right is 0.7648 (=1-BINOM.DIST(2,5,0.65,TRUE)), against 0.65 for one person; with nine voters it would be 0.8283. But if members are each right only 45% of the time, a five-person majority is right just 0.4069 of the time, and discussion before voting makes votes less independent. The team therefore keeps majority voting for low-stakes choices and uses consult-then-sponsor-decides for clinical protocol changes.

Worked example B (half solved): the second survey round

Six weeks later the same team repeated the survey.

1. Fill the gaps. [ ]
2. Which dimensions are still flagged under the agreed rule? [ ]
3. Which dimension improved most, and which action from round 1 most plausibly explains it? [ ]
4. Give one reason why the improvement might be overstated. [ ]

Exercise C (unsolved): a store opening team

Northfield Outdoor Co. formed a five-person team to open a new store in eight weeks. The psychological-safety scale has three items, scored 1 to 7.

1. Compute each respondent's scale score, then the team mean, SD, share favourable and $r_{wg}$.
2. Apply the flag rule from Worked example A. Is the dimension flagged, and for which reason?
3. The team has five members; how many communication links does it have, and how many would a team of eight have?
4. The store manager wants to decide the opening layout by unanimous vote. Give one advantage and two risks.
5. Week four is the midpoint. Design a 45-minute midpoint review agenda using Gersick's finding.

Connections to other cards

Common mistakes and misconceptions

1. Reading only the mean. Low agreement can matter more than a low mean; one member's very different experience is information.
2. Trying to find out who answered what. It destroys the safety the survey measures.
3. Treating psychological safety as being nice. It is the safety to disagree and admit errors; standards stay high.
4. Believing all conflict is good or all conflict is bad. Separate task, relationship and process conflict and handle each differently.
5. Assuming Tuckman's stages are a fixed timetable. They describe tendencies; deadline teams often follow Gersick's midpoint pattern.
6. Equating cohesion with performance. Cohesive teams are more vulnerable to groupthink if they are insulated and led directively.
7. Adding people to a late project to go faster. Links grow with the square of team size; coordination costs rise.
8. Using majority vote after an open discussion led by the senior person. Votes are no longer independent.
9. Over-reading tiny samples. With six respondents, a change of 0.3 on a 7-point scale may be noise.
10. Blaming individuals for loafing that the design invites. Make contributions identifiable and meaningful first.

Self-test questions

1. Name Tuckman's five stages (with Tuckman and Jensen 1977) and say what Gersick (1988) found instead.
2. Define team psychological safety as Edmondson (1999) did, and name the link she found to performance.
3. List Hackman's (2002) five conditions for team effectiveness.
4. Give four symptoms of groupthink and two remedies.
5. What did Stasser and Titus (1985) find about shared and unshared information?
6. Distinguish task, relationship and process conflict, and summarise De Dreu and Weingart (2003).
7. How many communication links does a 12-person team have? Give the Excel formula.
8. Five independent members are each right with probability 0.65. What is the probability that the majority is right? Give the Excel formula.
9. A dimension has scale scores 6, 6, 2, 6, 6, 6. Compute the mean and $r_{wg}$, and say what you would do.
10. Three conditions reduce social loafing according to Karau and Williams (1993). Name them.
11. When would you choose consensus over a leader decision after consultation?
12. A six-person team has a potential output of 10 units per person per week and produces 42. What is the process loss in percent, and name two likely causes?

My notes and next steps

Sources and further reading

• Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63(6), 384-399.
• Tuckman, B. W., and Jensen, M. A. C. (1977). Stages of small-group development revisited. Group and Organization Studies, 2(4), 419-427.
• 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.
• Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350-383.
• Nembhard, I. M., and Edmondson, A. C. (2006). Making it safe: the effects of leader inclusiveness and professional status on psychological safety and improvement efforts in health care teams. Journal of Organizational Behavior, 27(7), 941-966.
• Hackman, J. R. (2002). Leading Teams: Setting the Stage for Great Performances. Harvard Business School Press.
• Wageman, R., Hackman, J. R., and Lehman, E. (2005). Team Diagnostic Survey: development of an instrument. Journal of Applied Behavioral Science, 41(4), 373-398.
• Janis, I. L. (1972). Victims of Groupthink. Houghton Mifflin.
• Stasser, G., and 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.
• Latané, B., Williams, K., and 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.
• Karau, S. J., and Williams, K. D. (1993). Social loafing: a meta-analytic review and theoretical integration. Journal of Personality and Social Psychology, 65(4), 681-706.
• Jehn, K. A. (1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40(2), 256-282.
• Jehn, K. A., and Mannix, E. A. (2001). The dynamic nature of conflict: a longitudinal study of intragroup conflict and group performance. Academy of Management Journal, 44(2), 238-251.
• De Dreu, C. K. W., and 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.
• Rahim, M. A. (1983). A measure of styles of handling interpersonal conflict. Academy of Management Journal, 26(2), 368-376.
• Steiner, I. D. (1972). Group Process and Productivity. Academic Press.
• Vroom, V. H., and Yetton, P. W. (1973). Leadership and Decision-Making. University of Pittsburgh Press.
• Delbecq, A. L., and Van de Ven, A. H. (1971). A group process model for problem identification and program planning. Journal of Applied Behavioral Science, 7(4), 466-492.
• Schweiger, D. M., Sandberg, W. R., and Ragan, J. W. (1986). Group approaches for improving strategic decision making: a comparative analysis of dialectical inquiry, devil's advocacy, and consensus. Academy of Management Journal, 29(1), 51-71.
• James, L. R., Demaree, R. G., and Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69(1), 85-98.
• Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., and Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups. Science, 330(6004), 686-688.
• Condorcet, Marquis de (1785). Essai sur l'application de l'analyse à la probabilité des décisions rendues à la pluralité des voix. Imprimerie Royale.
• Textbooks and books, by name only: Thompson, L. L., Making the Team (Pearson); Edmondson, A. C., The Fearless Organization (Wiley, 2018); Robbins, S. P., and Judge, T. A., Organizational Behavior (Pearson).
• Open resource: OpenStax, Organizational Behavior (CC BY-NC-SA 4.0), https://openstax.org/details/books/organizational-behavior, chapters on groups and teams and on conflict.

Answer key

Worked example B.

1. Clarity: mean 5.17 (=AVERAGE(B3:G3)), $r_{wg}$ 0.97 (=1-VAR.S(B3:G3)/4), change +1.17. Task debate: SD 0.48 (=STDEV.S(B4:G4)). Relationship conflict: mean 2.02, change -0.48 (lower is better, so this is an improvement). Workload fairness: $r_{wg}$ 0.91.
2. None: every positive mean is at least 5.00, the reverse mean is 2.02, and the lowest $r_{wg}$ is 0.898 (psychological safety), about 0.90.
3. Clarity of direction, +1.17, most plausibly from the sponsor's restated, measurable goal with weekly data.
4. Any one of: people answer more positively once they know results are discussed (reactivity); six respondents give unstable estimates; the team may have learned which answers trigger a discussion; the member who scored lowest might be the one who now answers differently for reasons outside the team.

Exercise C. (1) Scale scores 5.67, 4.33, 5.67, 2.67, 5.33 (=AVERAGE(B2:D2)); mean 4.73, SD 1.28, favourable 60% (3 of 5 at 5 or above), $r_{wg}$ = 1 - 1.28^2 / 4 = 0.59. (2) Flagged for low agreement only: the mean of 4.73 is above 4.5, but $r_{wg}$ 0.59 is below 0.70. (3) 10 links for five people (=5*4/2), 28 for eight (=8*7/2). (4) Advantage: everyone is committed to the layout. Risks: deadlock and delay against an eight-week deadline; the most stubborn member controls the result; pressure to conform. (5) Example agenda: 5 minutes restate the goal and deadline; 10 minutes what is working and what is not (silent writing first); 10 minutes survey results; 15 minutes decide what to change in the plan, roles and norms; 5 minutes owners and dates.

Self-test.

1. Forming, storming, norming, performing, adjourning. Gersick found deadline teams keep their first framework until about the midpoint, then change sharply, rather than moving through stages.
2. A shared belief that the team is safe for interpersonal risk taking. Safety predicted learning behaviour, which in turn predicted team performance.
3. A real team, a compelling direction, an enabling structure, a supportive organisational context, expert coaching.
4. Any four of: illusion of invulnerability, collective rationalisation, belief in the group's morality, stereotypes of outsiders, pressure on dissenters, self-censorship, illusion of unanimity, mindguards. Remedies: leader states no preference first, devil's advocate, outside experts, subgroups, a second-chance meeting.
5. Groups discussed mainly information that all members held before the meeting and under-used information held by only one member, which led to worse decisions when the unshared information pointed to the best option.
6. Task: about the work; relationship: personal friction; process: roles, logistics, who does what. The meta-analysis found both task and relationship conflict negatively related, on average, to team performance and to member satisfaction; any benefit of task conflict depends on conditions such as trust and psychological safety.
7. 66 links (=12*11/2).
8. 0.7648 (=1-BINOM.DIST(2,5,0.65,TRUE)).
9. Mean 5.33 (=AVERAGE(6,6,2,6,6,6)); variance 2.67, so $r_{wg}$ = 1 - 2.67 / 4 = 0.33. The mean passes, but agreement is very low: one member has a very different experience. Discuss the pattern as a team and offer one-to-one conversations without trying to identify the respondent.
10. Identifiable individual contributions, a task that is meaningful to members, a group that members value (also: small groups, unique contributions).
11. When implementation depends on every member's commitment, members hold information the leader lacks and there is time; not in an emergency.
12. Potential 60, actual 42, loss 18 units or 30.0% (=(60-42)/60). Likely causes: coordination losses (handoffs, waiting, unclear roles) and motivation losses (social loafing).

Revision history

Standard and licence note

Based on the original research on team development, psychological safety, team effectiveness, group decision making and intragroup conflict cited above (Tuckman 1965; Gersick 1988; Edmondson 1999; Hackman 2002; Janis 1972; Jehn 1995; James, Demaree and Wolf 1984). The text, survey items, worked examples, exercises and questions are original; no published survey instrument, textbook text or case is reproduced, and Tallow Creek Medical Center and Northfield Outdoor Co. are fictional. This card is independent and not affiliated with or endorsed by any business school, publisher or professional body. Check your course's readings and any survey policy at your school or employer before you rely on it. Template licence: MIT.