Facilitating team communication and problem solving is challenging, but it is hugely important and incredibly rewarding. Successful teams not only produce high quality outcomes, the social processes that teams use in carrying out their work can enhance members’ capability to work together interdependently in the future, and the team experience itself can also contribute positively to the learning and personal wellbeing of individual team members (Hackman & Woolley, 2008).

However, as noted by Hackman & Woolley (2008), team success hinges upon our ability to create real teams that have a clear and compelling direction or purpose – teams that are composed of the right number and mix of members, such that they can address the key challenge or problem that is central to their purpose. Teams also need a supportive organizational context, a well-designed team task, and some quality team-focused coaching or facilitation.  Team success is also enabled by task-appropriate team norms, and while I have highlighted the norm of freedom as non-domination in previous blog posts, it should be noted that this higher-order principle of freedom as non-domination allows for flexible design of task-specific behavioural norms that support different work processes. In other words, the team needs to be free to creatively design the rules or norms that shape their work. They can do this better when working in a supportive organizational context with access to quality team-focused coaching or facilitation.  However, with freedom comes responsibility, and in efforts to maximize successful workings in the group, teams need to learn and understand many things.

Previously, I highlighted a variety of approaches to idea generation in a team context, and some factors that influence productivity gains and losses.  For example, the number and diversity of unique, clear, quality ideas generated will vary depending on the method of idea generation a team uses.  Learning about different methods and specific norms or practices that influence team behaviour is important. Teams need to experiment with different methods and creatively apply (and evaluate the impact of) specific norms or practices.  For example, what happens when they generate ideas without judgement, focus initially on exploration and not decision-making, generate ideas simultaneously before taking turns to list and clarify, cycle between independent and interdependent idea generation strategies, and so on, etc.  Of course, in addition to generating ideas, there are many other tasks that teams will engage in.  Different psychological and social factors will influence performance on different tasks in different ways, and different methods and norms of behaviour will be needed to maximize successful workings for the team, depending on the task.

For instance, a team in collaboration with a team facilitator might distinguish between four types of tasks and consider the implications for their work: (1) Generate tasks, (2) Negotiate tasks, (3) Choose tasks or decision-making tasks, and (4) Execute tasks (McGrath, 1984; see Woolley, Aggarwal, & Malone, 2015).  As noted, generate tasks involve idea generation, relevant for example when describing the range of problems in a problem situation, or when generating options in response to problems. Negotiate tasks involve perspective taking, resolving conflicts of interest or viewpoints. When a team works to resolve societal problems they may face a variety of challenges associated with negotiating meaning, clarifying ideas, resolving conflict related to team activities, balancing perspectives as to the nature of evidence relevant to a systems model they are developing, and so on.  While facilitation during generate tasks requires methods and practices designed to promote diversity of unique, quality ideas and the clear communication of ideas, facilitating negotiate tasks involves practices that foster perspective taking, resolution of conflict, reinforcement of key values, norms, or rules, reorientation to core group goals, and so on.

Choose tasks or decision-making tasks involve team members selecting from a range of alternatives. When selecting options in response to societal problems, this involves decision-making or choice selection derived from a judgement as to the ‘best’ choice, as often there is no ‘correct’ choice. Facilitation in this context involves the promotion of critical and reflective thinking, weighing up available evidence and drawing conclusions or judgements that influence key choices and decisions. Choose tasks or decision-making tasks often involve establishing criteria for choices or decisions, for example, choosing from a range of alternatives based on an estimate of their potential impact, feasibility, and the potential of the team to implement key choices as actions, given current resources and capabilities.

Finally, execute tasks involve teams acting directly upon the world in one way or another.  At the team level, this requires action coordination and distributed skill to successfully enact specific choices, options, or solutions.  John Warfield, in developing his approach to applied systems science, described the science of action as a central pillar that builds upon and helps to realise coordinated actions that emerge as outputs from the science of design and complexity. Understanding the complexities of societal problems and designing systems models that capture this complexity is one thing, but implementing coordinated actions that help resolve these problems is another thing completely.  Indeed, intelligence without action is inert, and in the context of applied systems science, action without intelligence is ill-informed.  More generally, collective intelligence implies the coordination of individual actions, and one might argue that every form of collective intelligence described here, for every collective ‘task’ described, involves some form of coordinated execution of task actions. Certainly, abstracted as a principle, coordination is critical for all forms of intelligent action.  Empirically, laboratory studies that have sought to measure the ‘general’ collective intelligence of groups (i.e., by extracting a common statistical ‘general intelligence’ factor that measures the ability of teams to solve a diverse range of problems), report that ‘general’ collective intelligence of groups is related to the groups’ ability to coordinate their activity (Aggarwal, Woolley, Chabris, and Malone, 2015). Also, at the applied project level, in efforts to resolve societal problems, when multiple coordinated actions derived from collective intelligence work are translated into actions on the ground, the team involved in the collective intelligence work can be instrumental in coordinating the societal projects.  I’ll provide some examples of this in future blog posts.

As such, teams have a lot to learn about coordinating their activity in the context of different tasks.  Collective intelligence requires contextual awareness of the many factors that influence their coordinated action in context. Returning now to the social context, let’s consider some of the dynamics at play. As noted by Strauss et al (2009) groups may have a variety of social preferences or inclinations that influence their collective intelligence. For example, for the ‘smooth’ functioning of the social group, they may prefer to work in more homogeneous teams that include like-minded people, as opposed to more heterogeneous teams where there is greater diversity amongst members; they may have an inclination to discuss shared (common) knowledge as opposed to both shared and unshared (unique) knowledge that is distributed amongst team members; they may also have an inclination to suppress divergent information to avoid ‘standing out’ as unique or different, and ultimately to avoid social exclusion because they are ‘different’; they may prefer consensus to dissent and conflict; and they may be uncomfortable dealing with uncertainty as regards the nature of evidence or the choices amongst alternatives. Efforts to push against these inclinations or preferences may produce problems that the team facilitator needs to anticipate. For example, efforts to increase cognitive diversity in the team may backfire and may result in reduced cohesion, greater conflict, and reduced team member satisfaction (see Mannix & Neale, 2005). The facilitator and the team need to anticipate these issues and be prepared to work in ways that maximize their success.  There are many metaphors we could use to describe the work of the facilitator in this context: in some ways, the facilitator is like a chess player, anticipating in advance the impact of every team move; in other ways, the facilitator is like the conductor of an orchestra, coordinating activities as they emerge and change over time.

Whether the activity focus of a team is on idea generation, negotiation, or deliberation in advance of choice making, a primary issue to consider in a social context is what type of information people have a preference for sharing. A facilitator can only coordinate what is shared, and thus may need to prompt team members to share their unique knowledge.  In dynamic conversational exchanges, it can be difficult to overcome the common knowledge effect. Specifically, teams will often focus on information or knowledge that is commonly held among all group members (see Strauss et al., 2009). Therefore, unique (and relevant) information held by individuals is not always shared with the team. When it is shared, it may be overlooked as it is overshadowed by the information or knowledge that ‘everyone is aware of’. For example, a team addressing the problem of unemployment and homelessness in their city may focus on and discuss what they all know about the problem of unemployment and homelessness, rather than sharing, discussing, aggregating and evaluating both their common knowledge and all the unique knowledge that is known only to individual team members, or a smaller sub-group of specialists in the team. Unfortunately, when only common knowledge is shared and discussed, team judgements and decisions discount potentially important information – information that would result in a different and potentially better judgement and decision, if it were shared. As reviewed by Kerr and Tindale (2004), teams may find it difficult to overcome the common knowledge effect for a variety of reasons: as everyone in the team is aware of common knowledge, research indicates that it is more likely to be discussed, especially early in the discussion. Also, where there is need to reach consensus – and often there is pressure to do this quickly –in a situation where most of the members already share the same preference, this can lead to reduced information exchange and early consensus.  Research also indicates that people prefer to present and receive information that is shared amongst team members, and, somewhat surprisingly, people are perceived as more knowledgeable, competent and credible when they share information that others already know. Sharing common knowledge may be deemed more socially acceptable than sharing one’s unique knowledge, as the social norm may be to ‘not stand out’ from the crowd. As noted by Kerr and Tindale (2004), team members may enter meetings with preferences supported by common knowledge and, even when shared, information inconsistent with these preferences may be discounted or misinterpreted.  As such, people may be resistant to changing their initial judgment, even in a situation where new knowledge is presented suggesting that their initial judgement is false.

Facilitation in this context can be challenging. If we return to the orchestra metaphor, in the extreme case, the exclusive sharing of common knowledge reflects a scenario where everyone is playing the same instrument and the same sequence of notes. It’s unlikely to be Beethoven or Mozart they are playing. For teams focused on societal problem, it’s unlikely to be a comprehensive understanding of a societal problem they produce. The facilitator needs to prompt a different production. In situations where common knowledge fosters common attitudes within individual team members, social interaction and discussion amongst team members may foster the emergence of increasingly extreme attitudes, as everyone appears to agree with one another and this serves to reinforce and strengthen their common knowledge and attitude (Isenberg, 1986). This phenomenon is known as group polarization and it highlights a tendency for relatively homogeneous groups to make more extreme decisions after discussion than would be expected based on a simple average of their pre-discussion views.  As such, a facilitator may anticipate a potential scenario unfolding – not uncommon within political groups or business or organisational contexts — where individuals are exposed to largely common and biased information in advance of meeting to work to resolve a shared problem, and they then conduct a poorly facilitated ‘brainstorming’ session with a predominant focus on commonly held ideas, thus emerging with a more extreme, biased, and erroneous judgement in relation to the problem and proposed solutions.

Even in situations where arguments shared amongst team members are deeply flawed by reference to common standards of reflective thinking, group polarization may occur in part because people often agree with, and are persuaded by, others who share common beliefs and attitudes.  As noted by Strauss et al (2009), this process of agreement and persuasion may be further compounded by common inclinations toward confirmation bias, that is, the tendency for individuals and groups to seek out and focus on information or evidence that supports a pre-existing belief or hypotheses. Although Karl Popper famously proposed that a key indicator that a good theory is that it is open to being disconfirmed, people often become attached to their view and seek out information or evidence that confirms rather than disconfirms their theory.  The challenge is to design and facilitate teams such that they freely share their unique knowledge, develop complex and increasingly valid (as opposed to polarized and increasingly biased) knowledge structures, and treat their knowledge structures as theories that are open to disconfirmation and revision.

Certain things can help and may be open to design and advance planning on the part of the team facilitator. For example, building teams that include members with a greater diversity of opinions may inhibit tendencies toward confirmation bias, and may be better than prompting select members within more homogenous teams to play “devil’s advocate” (Schulz-Hardt, Jochims, and Frey, 2002). But even when working with a diverse team, within the meeting itself, the team facilitator need to help the group select appropriate methods and adopt a reflective stance, promoting curiosity, creativity, and critical thinking – thus helping a team to move beyond narrow, biased lines of enquiry.  A team facilitator, in adopting a reflective stance, also helps a group to reflect honestly on their current state of knowledge and avoid another common pitfall of teams — overconfidence in their opinions and judgements.  Much like group polarisation and confirmation bias, overconfidence can arise when team members repeatedly confirm one another’s opinions, or when common, hypothesis-confirming arguments are shared among group members. While individuals may have a tendency to overestimate and overrate their knowledge, judgement and performance, this overconfidence effect can be even stronger in teams, particularly in situations of uncertainty where there is no ‘right’ or ‘wrong’ answer (Zarnoth and Sneizek, 1997; Sneizek, 1992).

Notably, for many complex social problems there is no simple ‘right’ or ‘wrong’ answer as regards the ‘single best’ course of action to resolve the problem.  Indeed, it is more often the case that multiple complementary and coordinated actions are needed – actions that match the complexity of the problem. Unfortunately, polarized groups sharing common knowledge and seeking only confirmatory evidence to support their proposed solution to a problem may ‘dig their heels in’ and maintain a very confidence stance in relation to their representation of the problem and proposed solution to the problem. Indeed, both more homogeneous groups that lack diversity of options, and groups who are demonstrably wrong in their judgement, may exhibit higher confidence than more diverse, higher performing groups — their unrealistic confidence may indeed worsen confirmation bias tendencies and make the group less likely to seek out evidence, information, or opinions from others (Strauss et al., 2009). Also, while increasing group size may enhance performance on tasks where there is a definitive correct answer (e.g., computation tasks), increasing group size tends to increase confidence, but not accuracy, in tasks that require reflection and judgement (e.g., forecasting tasks; Zarnoth and Sneizek, 1997).

The exercise of reflective judgement is necessary and should be established as a norm, particularly if the team seeks to move from facilitator-led to a peer-led working. This implies that team members exercise of a range of critical thinking dispositions.  Carefully addressing the ideas of sceptical individuals, or those who doubt the common, overconfident judgement of the majority is an important task for the facilitator.  The facilitator acts as a conduit for the emerging collective judgement of the group and must model and maintain a curious, reflective, and neutral stance in the process.  While a good facilitator will often maintain a humble stance when exercising their reflective judgement in a collective intelligence session, they must nevertheless understand the key skills and dispositions of good critical thinkers and both exercise these skills and (gently and strongly) facilitate team members to do the same. In addition to the chess playing and orchestra conducting metaphors, we can add another potentially useful metaphor: as the Tai Chi masters might describe it, the facilitator is like steel wrapped in cotton. The goal is to steel the group against potential process biases and steer them softly and strongly in a direction that maximizes their collective intelligence. When facilitating a group to address a societal problem, the facilitator steers the process with a curious, reflective, and neutral stance, but does not generate content relevant to the problem – the team provide the content knowledge that feeds into their representation of the problem, their proposed solutions to the problem, and their action plan moving forward.

Notably, without a facilitator present to uphold sufficient humility and maintain a curious, reflective, and neutral stance in relation to the flow of communication in the room, protecting dissenters who present critical, relevant and valid evidence that bears upon a judgement process, it is possible that group members may put significant pressure on dissenters to change their view.  Indeed, individual group members may be rejected by the group if they do not change their view and this social exclusion or rejection can cause significant psychological distress. This behaviour clearly violates the principle of freedom as non-domination and one task of the facilitator is to uphold this principle in action. However, the pressure to conform and move toward uniformity of opinion can be intense in a group performance context.  Also, while freedom as non-domination is an ideal that we have espoused, and implies that everyone should have equal power and influence in the teamwork context, in reality, where clear dominance hierarchies exist, the pressure to conform can be intensely dominating, particularly for low-status group members.  Unless the facilitator can cultivate and maintain freedom as non-domination in context, lower status team members, when compared with higher status team members will have a tendency to seek social acceptance, focus on shared information, speak less and have less influence, which can result in mixed-status groups sharing less unique information and arriving at poorer decisions than do equal-status groups (Strauss et al., 2009).

When individual team members seek acceptance, and when the group pushes for conformity, uniformity, and a strong sense of cohesion (or ‘we’ feeling), the effect on critical and reflective judgement can be very negative (Janis, 1972, 1982).  Rather than exploring a problem comprehensively, carefully, and reflectively, groups that push for cohesion and uniformity may suffer a phenomenon known as groupthink.  A strong sense of cohesion (or ‘we’ feeling) may fuel a tendency amongst individuals to agree with one another, or engage in concurrence seeking behaviours.  By analysing a number of real-world cases of group decision making, Janis (1972, 1982) proposed that groupthink can arise in situations of high group stress or in situations where there is pressure on the group to arrive at a solution to a problem rapidly. Another catalyst of groupthink, says Janis, is the presence of a directive leader, who enters the group problem solving situation with a strong set of preferences, and dominates other group members in pushing for agreement. Furthermore, if the group is insulated from outside expert opinion and has no established norms or methods for systematically searching and evaluating information and evidence, then group member may simply ‘agree’ with the leaders’ proposals.  Notably, although there is no basis for doing so, a directive leader may ‘infer’ the agreement of team members from their silence.  If, says Janis, group members have little confidence that solutions other than that proposed by the leader are viable, and if specific group members take on the role of suppressing dissent and ensuring agreement, there is little hope that critical and reflective thinking will prevail in the group.

Naturally, group cohesiveness can take on a variety of different forms and cohesiveness is not always negative.  For example, a strong sense of cohesion can derive from a shared purpose and shared commitment to resolving a problem, or a shared embrace of common norms of behaviour (e.g., “we embrace the principle and practice of freedom as non-domination”). But this is very different from cohesiveness that arises from a strong desire to simply ‘get along’ or ‘like’ one another.  It is also very different from cohesiveness derived from a strong desire in advance of meeting to agree about the facts of a problematic situation. Naturally, it is also very different from cohesiveness derived from shared agreement as to options we will pursue ‘in advance of’ systematically gathering the facts and exercising good judgement.

The norm of freedom as non-domination, by default, liberates team members from conformity pressure, a need to gain social acceptance, pressure to share common knowledge only, or pressure to defer to the option of a leader.  Freedom as non-domination reinforces collective leadership, it fosters cohesiveness grounded in shared freedom and shared power, and it allows for openness and the sharing of non-common or unique ideas.  Freedom as non-domination affords everyone the opportunity to offer a ‘dissenting’ view.  More simply put, freedom as non-domination allows one to speak one’s mind and exercise one’s full creative, critical and reflective potential in a group problem solving context.  Freedom as non-domination liberates perception and thinking and allows one to identify, generate, share, integrate, and apply information, including information that might have previously been classified as dissenting information, and this ultimately enhances the performance of the group. In this new context, ‘dissent’ simply becomes part of everyday thinking, enquiry, a normal part of collective intelligence. Why should it be otherwise?

As the principle and practice of freedom as non-domination becomes more commonplace within educational, organisational, and political systems, and society more generally, groups can plan and implement collective intelligence projects with greater ease and forethought.  The historically and empirically observed dynamics of team performance across a variety of contexts can enter awareness and become part of the context shaping the ongoing design and facilitation of coordinated team efforts in response to societal problems. This implies a new form of mindful awareness, awareness needed for the facilitation of teams. The scientist, chess player, orchestra conductor, and Tai Chi practitioner become one.

References

Aggarwal, I., Woolley, A. W., Chabris, C. F., and Malone, T. W. Cognitive diversity, collective intelligence, and learning in teams. Proceedings of Collective Intelligence 2015, Santa Clara, CA, June 2, 2015

Hackman JR, Woolley AW. (2008). Creating and leading analytic teams. In: Rees RL, Harris JW A Handbook of the Psychology of Intelligence Analysis: The Human Factor. Burlington, MA: Centra.

Janis, I. L. (1972). Victims of groupthink. Boston: Houghton Mifflin.

Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes (2 ed.). Boston: Houghton Mifflin.

Kerr, N. L., & Tindale, R. S. (2004). Group performance and decision making. Annual Review of Psychology, 55, 623-655.

Mannix, E., & Neale, M. A. (2005). What Differences Make a Difference?: The Promise and Reality of Diverse Teams in Organizations. Psychological Science in the Public Interest, 6(2), 31-55.

McGrath, J. E. (1984).  Groups: Interaction and Performance. Prentice-Hall.

Schulz-Hardt, S., Jochims, M., & Frey, D. (2002). Productive conflict in group decision making: Genuine and contrived dissent as strategies to counteract biased information seeking. Organizational Behavior and Human Decision Processes, 88, 563-586.

Sniezek, J. A. (1992). Groups under uncertainty: An examination of confidence in group decision making. Organizational Behavior and Human Decision Processes, 52(1), 124-155.

Straus, S. G., Andrew M. P., James B. B. and Jacob W. D. (2009). The Group Matters: A Review of the Effects of Group Interaction on Processes and Outcomes in Analytic Teams. Santa Monica, CA: RAND Corporation.

Woolley, A. W., Aggarwal, I., & Malone, T. W. (2015). Collective intelligence in teams and organizations. In T. W. Malone & M. S. Bernstein (Eds.), The handbook of collective intelligence. Cambridge, MA: MIT Press

Zarnoth, P, & Sniezek, J. A. (1997). The social influence of confidence in group decision making. Journal of Experimental Social Psychology, 33, 345-366.