Part 2: What makes teams successful?


Last week I wrote a post about research from Michael Klug and James Bagrow based on mathematical analysis of Github data regarding ‘high impact’ teams (see Part 1: What makes teams successful?). The skinny?

They found that high impact teams — those that the community have high regard for — share several key attributes:

    • High-impact teams tends to be larger than small teams. However, Github teams tend to be small, with less than one percent having more than 10 members.
    • High-impact teams are more focused than lower-impact teams of the same size, and are likely to have members with diverse experience.
    • Even in larger teams, high-impact teams have core and support cadres. A small number — sometimes just one — of the team do the majority of the work, while other, non-core members act in support roles.
    • Perhaps just as important: high-impact teams are more likely to have members that are core members of other teams.
    • High-impact teams are more likely to be ‘dominated’: where the lead member contributed more work than all the other contributors combined.

    As the authors state,

    This mixture of size, focus, experience, and diversity points to underlying mechanisms that can be used to maximize the success of collaborative ventures.

    Other research on team interaction sheds light on the role of the the strong ties in collaborative teams. Research publish in June by Yves-Alexandre de Montjoye, Arkadiusz Stopczynski, Erez Shmueli, Alex Pentland, and Sune Lehmann, examined the way that teams’ networks work. They found that

    while an assigned team’s performance is strongly correlated with its networks of expressive and instrumental ties, only the strongest ties in both networks have an effect on performance. Both networks of strong ties explain more of the variance than other factors, such as measured or self-evaluated technical competencies, or the personalities of the team members. In fact, the inclusion of the network of strong ties renders these factors non-significant in the statistical analysis.

    The research was conducted using 45 teams of four students during a semester at MIT. Eighty participants worked in teams for three separate projects for one course, and no one worked with the same person twice. The performance was measured by the grades given by the instructors to the team reports.

    The authors use J.R. Lincoln’s distinction between expressive and instrumental ties. Expressive ties are those that occur between friends while instrumental ties are those that arise in professional settings, and which support information transfer, esprit de corps, and interpersonal familiarity. The researchers used a questionnaire to determine the degree of connection between the students.

    The results show that those that already were friends prior to the project work had the best results:

    Does a dense network structure help a team to perform well or does a performing team create dense networks? In this study, we measured the network of expressive ties before the experiment started. We then assigned participants to teams, and we see a positive correlation between the strongest expressive ties and team performance.

    The work done by the teams required complex problems that required creativity and ‘applying gained knowledge in a novel context’. Much like the work most of us are confronted with at work. The researchers also found that the strength of ties in instrumental and expressive networks is not independent, and that instrument tie strength was highest in the highest strength of expressive ties, as well.

    Based on this research, a few comments on teamwork. Businesses that would like to see higher performing teams would be wise to be aware that expressive ties are the source of the best outcomes. The obvious comment is that this research supports the principle of self-forming teams. When people are allowed to choose who to work with they will likely opt for trusted friends and contacts. Note that this does not mean that all teams will have uniform performance, but on average, people working with friends will have higher performance than they would with non-friends.

    Comments are closed.