A team must learn in order to get better at what it does, produce more work, and/or lower costs. To learn, a team must display learning behaviors such as experimenting with new ideas, communicating well, and writing down knowledge in a reusable form. We know that people who are similar are more willing to share ideas and tend to communicate better, but diversity contributes outside perspectives. So what is the best balance of similarity and diversity for team learning?
To find out, researchers conducted surveys of 724 individuals from around the world on 156 teams of various types in the pharmaceutical and medical device industries. Teams with good learning practices were very diverse as a whole, but in a way that moderately similar individuals could form natural subgroups. For example, consider both Team 1, a sales team of middle-aged Asian women on the job for a few years; and Team 2, each member of whom is very different from each other in job skills, age, experience, ethnicity, etc. Most teams have a level of diversity between these extreme examples, so two or more members share some—and only some— characteristics (personal and/or job-related.) For example, a CQI team from a major corporation could have: young gaming engineers, older male support staff, highly educated Japanese managers, etc. That team has the advantages of diversity, but each member feels safer expressing ideas because someone like the member is in the room. As the authors write, "Similarities within subgroups…enable information and insights to surface, while differences across subgroups ensure that a diversity of insights is considered." But if subgroup members are too similar, the advantages of the teams' diversity may disappear…the team may break down into nondiverse subteams.
A team also practiced more learning behaviors if it felt empowered, was held accountable for performance by its manager, and was supplied by its company with a knowledge management system. But again, a team whose subgroups (if any) shared very few or very many characteristics were not helped as much by these supporting variables as those with subgroups that shared a moderate number of characteristics.
"The implication," the authors write, "is that leaders may have to behave differently toward the teams they manage depending on their composition in terms of subgroup strength." In particular, "if a team has weak or strong subgroups, external leaders can stimulate learning by engaging in performance management." In the study, "performance management" meant the team's manager encouraged it to:
Editor's Note: A weakness in the study design is that the surveys did not test whether subgroups actually formed—that is, whether people behaved within subteams in the way the authors assume they would. But their assumption fits the research on similarity.
Source: Gibson, C., and F. Vermeulen (03), "A Healthy Divide: Subgroups as a Stimulus for Team Learning Behavior," Administrative Science Quarterly 48(2):202.