When Does Diversity Make a Team Stronger?
Background
In firms, in schools, in sports, the diversity of a team’s members is said to matter a great deal. Are teams stronger when they bring different people together, or when their members resemble one another?
The answer obviously depends on the kind of task the team performs and on what sort of diversity is meant: diversity of skill types, of skill levels, of personality, or of age, gender, and ethnicity.
This paper focuses on two different kinds of diversity. One is ability, how efficiently a member can carry out the same work. The other is relational orientation, how much a member treats a teammate’s outcome as her own concern, which is also referred to as social preferences. When and how do these two kinds of diversity raise team performance?
The answer is not simple. Diversity raises performance when it lets each member specialize in what she does well, and lowers it when it disrupts the coordination that joint work requires. Nor is there any guarantee that ability diversity and relational-orientation diversity push team performance in the same direction. Organizational research has documented a tendency for diversity in expertise and ability to help performance while diversity in values and social orientation tends to hurt it, and the explanation has usually been sought in interpersonal friction. This paper shows that the same pattern can arise from the members’ strategic effort choices alone, with no such friction anywhere in the model, using a game-theoretic analysis.
What the paper does
Consider a simple model of team production with two members. The two supply effort to produce a joint output; effort is privately costly, and an abler member carries out the same work more efficiently. Members also differ in how much they care about a teammate’s payoff, which is their relational orientation.
The key is the nature of the task. Tasks that require close joint work make the two efforts complements, so that each member’s contribution is worth more when the partner also contributes. Divisible routine work makes them substitutes, so that one member’s effort can stand in for the other’s. Holding the team’s averages fixed, the paper varies only the dispersion of ability and of relational orientation and asks what happens to team performance. Both horizontal teams, in which the two move simultaneously, and vertical teams, in which one leads and the other follows, are analyzed.
What it finds
Diverse hands, aligned hearts
The central result is that the two kinds of diversity work through separate channels. Ability diversity raises performance whatever the task, because effort can be concentrated in the abler, lower-cost member. Relational-orientation diversity, by contrast, depends on the task: it lowers performance in closely coordinated work, where efforts are complements, and raises it in divisible work, where they are substitutes. Work that requires coordination and genuine teamwork therefore calls for a team with diverse hands but aligned hearts.
The reason can be put simply. Dispersion in ability raises the scale of the team’s total effort. Dispersion in relational orientation barely changes the total; it upsets the balance between the two members’ contributions. Imbalance is a loss in complementary tasks, where meshing matters, and a gain in substitutable tasks, where the contributions are interchangeable. Hence the reversal of sign.
Teams do better when the abler member is also the one who cares
Holding the ability gap and the orientation gap fixed, a team in which the abler member also has the higher relational orientation always outperforms the team in which the two are split. Members with a higher relational orientation supply more effort, so the team gains most when that member is also the abler one. Put the other way round, a perfectly homogeneous team is never optimal: whenever attributes can be reshuffled, performance can always be improved.
Leadership: who should move first
In a vertical team, where one member moves first, the question of who should lead arises. Ability diversity is not very sensitive to the assignment: it raises performance whoever leads. Relational orientation is different. What matters is which role the more prosocial member is placed in, and the optimal placement flips with the task: she should follow when efforts are complements and lead when they are substitutes. The logic is that the most responsive member belongs in whichever role tends to supply too little effort.
The paper also shows that observing and measuring members’ social preferences has value, because it makes better team pairings and better leadership choices possible and thereby raises team performance. That value grows the stronger the task interdependence, the larger the dispersion among members, and the more reliable the measurement.
Why it matters
Implications for team composition and design
Diversity is neither uniformly good nor uniformly bad. What matters is matching the shape of the dispersion to the nature of the task. In work requiring close coordination, such as R&D, joint research, integrated development teams, and team sports, diversify ability while keeping hearts aligned. In divisible routine work, both kinds of diversity help.
Implications for human resources and organization
Organizations routinely collect information on skills and experience. The implication of this paper is that information on social orientation carries a value of its own, distinct from that, because it improves the quality of decisions about who should be paired with whom and who should lead.
Testability and next steps
If relational orientation can be measured, the central predictions can be tested as they stand: greater dispersion in orientation should lower performance in complementary tasks and raise it in substitutable ones. I hope to take this to experimental testing in future work.
This research was supported by JSPS KAKENHI Grant Number 24K04912.