Social Interaction and Team Performance

Background

Many organizations invest heavily in team-building activities and workspace design intended to encourage social interaction. Google’s campuses are designed to produce chance encounters, and the Pixar headquarters designed by Steve Jobs is famous for a layout that maximizes spontaneous conversation.

Nike runs sporting events between teams throughout the year, and Airbnb has built a culture that encourages employees to share experiences. The LEGO Group holds regular informal events built around play and learning. These firms are all aiming to raise trust, belonging, innovation, and team performance by forming informal communities inside the workplace.

Despite these substantial investments, however, clear evidence has been lacking on how informal interaction between people at work improves team outcomes.

This study asks what psychological mechanism makes social interaction improve team performance, and in particular what mechanism operates when the task requires complementary effort from team members. We combine a game-theoretic model with a laboratory experiment that tests the predictions the model generates.

What we did

Our hypothesis was that social interaction between teammates narrows the differences in how much they care about each other, and that this is what raises team performance.

Interacting with someone who cares about you tends to increase the care you feel toward that person, while interacting with someone who does not tends to decrease it. Social interaction should therefore let teammates read each other’s feelings toward them, and this should shrink the actual differences in mutual regard within the team. We call this narrowing a reduction in the prosociality gap.

Our hypothesis is that such a reduction in the prosociality gap improves performance on team tasks where cooperation is indispensable. The study tests this both with a mathematical model and with a laboratory experiment on university students.

The theoretical framework

We build a model of team production in which the members’ efforts are complementary, and derive predictions about how social interaction that reduces the prosociality gap, the difference in how much teammates care about each other, affects individual and team performance.

The model predicts the following.

1. Social interaction necessarily increases the effort of the less prosocial member. Whether it increases the effort of the more prosocial member depends on effort complementarity.
2. Social interaction improves team performance regardless of the initial distribution of prosociality.
3. This performance effect is stronger in teams with greater effort complementarity and higher loss aversion.

The experiment

To test these predictions, we ran a pre-registered laboratory experiment with 74 two-person teams of university students. Teams were randomly assigned to one of two conditions.

• Treatment: the two members worked together on a 10-minute joint brainstorming exercise.
• Control: the same brainstorming exercise was done alone, without conversation with the teammate.

Every team then performed the SLAP typing task. This is a collaborative computer task in which one member is responsible for the “S” and “A” keys and the other for the “L” and “P” keys, and the pair enters the sequence “SLAP” as many times as possible. The task is designed to require a high level of complementary effort. High scores are rewarded, so each team cooperates in order to score well. The main test compares performance on this task between the treatment and control groups.

This randomized design lets us measure quantitatively how social interaction affects the prosociality gap within a team and team performance.

We also examined the role of emotional perceptiveness, the accuracy with which a member predicts how much her teammate cares about her, in order to explain why social interaction narrows the prosociality gap, and we looked at which kinds of teams gain most in performance from social interaction.

What we found

Our results strongly support the hypotheses derived from the theory.

1. Social interaction reduces the prosociality gap by 42%. Teams that did a short joint brainstorming exercise had a significantly smaller prosociality gap than control teams.

2. Social interaction improves team performance by 14.4%. Treatment teams achieved significantly higher scores on the SLAP task than control teams.

3. The reduction in the prosociality gap is the main reason social interaction raises performance. Mediation analysis shows a significant indirect effect: social interaction reduces the prosociality gap, and that reduction raises performance. About 38% of the performance gain came through the narrowing of the gap. Interestingly, no mediation was found through the level of closeness or the level of prosociality that members felt toward each other.

4. Emotional perceptiveness plays a central role in how social interaction narrows the prosociality gap. Perhaps most interesting of all, social interaction improves members’ ability to predict each other’s prosocial attitudes accurately, and this improvement in turn narrows the prosociality gap. This perceptual channel accounted for a remarkable 98% of the reduction in the gap.

5. The performance effect of social interaction was stronger in same-gender teams, especially female pairs, and in teams whose members were more loss averse.

Our conjecture is that communication barriers are lower in same-gender teams, so the effect of interaction is stronger; and that for people who fear failure, talking beforehand creates a psychologically safe environment, so the reduction in interpersonal risk matters more to them.

Why it matters

We believe these findings carry important implications for organizations.

1. Even brief, structured social interaction has a large effect. Ten minutes of joint brainstorming substantially improved subsequent team performance.

2. Emotional alignment is the key mechanism. Organizations often focus on building team cohesion in general. Our study suggests that what drives the performance gain is specifically the alignment of prosocial attitudes between members, together with the accuracy with which each member can infer how the other feels toward her. It may be worth including measures of prosociality and emotional perceptiveness in team development assessments.

3. Emotional perceptiveness can be improved through interaction. The ability to read a teammate’s intentions and attitudes accurately turns out to be a skill that structured interaction can develop. Team-building activities should be designed to raise emotional perceptiveness and prosociality alignment.

4. Team composition should be taken into account. The benefits of social interaction can differ with team characteristics such as gender composition and attitudes toward risk. Team composition should be considered when planning social-interaction initiatives, recognizing that some teams will gain more than others.

Our study provides evidence that social interaction reduces the prosociality gap by improving emotional perceptiveness, and that this improves team performance. We believe it gives scientific grounding to the intuitive belief that strengthening informal connections between people yields performance benefits, while also offering concrete guidance on how to maximize those benefits.