The news by Google on how to create a dream team stimulated some conversation in my academic circles. Especially the point that Google was not able to replicate the importance of team composition created some stir. For me, this was obvious. I never considered team composition to be an important factor predicting team performance. Of course, it makes a difference if you have experts and non-experts in your team, and one neurotic or extrovert team member can create quite a lot of troubles. But a team of experts does not make an expert team.
I haven’t met many people who dispute this. Most of my colleagues adhere to a social constructivist perspective on learning and human behavior. In short, performance depends on the environment. Individuals are not in a vacuum but are interacting with others. What a person learns, depends on the person and his/her interaction with ‘learning units’, other peers and learning material. In light of this, it is strange that most of my colleagues do not use social network analysis and consider it only a method.
So how does team composition relate to team performance? Of course it depends on the compositional factor. Frequent compositional aspects considered are personality of team members, their knowledge and their attitudes. The next aspect to consider is how those individual compositional factors aggregate at the team level. This is an important question, as teams can not have a personality. The personality of the team depends on the personality of team members. Often the personality of team members are averaged and this mean personality level is described as the team’s personality. However, you could also say that variance in personality is important. In other words it is not the average level, but the distribution of personality scores. These are important questions to consider when analyzing team composition.
Focusing on the mean level of compositional factors and its impact on team performance in field studies, the correlation between various factors and team performance ranges from 0.06 to 0.35 based on a meta-analysis . For example, team emotional stability (mean levels) and team performance are only correlated by 0.06. The team’s mean level of emotional stability thus has a negligible impact on team performance. However, team mean levels of collectivism and preference for team work have a much stronger influence on team performance (0.35 and 0.22 respectively). These two compositional factors thus have a moderate influence on team performance. Team members mean level of general mental ability has also only a moderate relationship to performance (0.18). Factors such as team members preference for team work contribute more to the team’s performance than team members cognitive abilities.
While this seems to support my claim that team composition doesn’t matter (that much), a good argument for my claim should also include results about how team level constructs, such as shared mental model, relate to team performance. A meta-analysis also exists for this relationship . This meta-analysis summarized the impact of various team factors on team cognition. Team cognition is an emergent team state that describes how the team is mentally organizing and accessing the knowledge important to achieve their goal. If we visualize all the behaviors, actions, and attitudes that happen within team members as a black box, team cognition is what comes out of this black box. Examples of team cognition are the team’s mental model, expertise distribution, or task coordination. The correlation between team cognition and team performance ranges is 0.31 for objective performance ratings and 0.44 for subjective performance ratings. Thus, it is similar to the strongest correlation between team composition variables and team performance.
Is my standpoint supported? Yes, but the foundation is less strong that I wished. I would have preferred a stronger correlation between team cognition and team performance. That would have given me more ground to stand on. To conclude, team composition is an important aspect to consider when assembling teams, but it only provides the starting point. How team members interact, and how they mentally organizing their knowledge plays as much of a role.