(Mis-)Perceptions of the Importance of Team Composition and Team Governance in Non-Routine Analytical Tasks
Dominik Grothe (LMU Munich); Florian Englmaier (LMU Munich); David Schindler (Tilburg University); Simeon Schudy (LMU Munich)
Abstract
Team-based organizational structures have become the building block of modern firms and scholars widely agree that teamwork is an important component of economic success and innovation. The most pressing question for managers (and scholars alike) is what renders teamwork in complex problems successful. Only when forming the right expectations about the main drivers of performance, teams can cost-effectively succeed. This study elicits expectations about the relative importance of team composition (team size, gender composition, experience) and team governance structures (monetary and non-monetary team incentives and leadership) for team performance in a non-routine task in a representative sample (n=1,500) using an incentivized discrete choice experiment. We exploit the unique opportunity to contrast these expectations with actual performance data from 1,358 teams working on the non-routine tasks for which the expectations were elicited. We collected the performance data in a series of large-scale field experiments in which the respective team characteristics and team governance structures varied. Our results reveal that participants correctly perceive gender diversity in teams and experienced team members as an important driver of team success and they (also correctly) expect monetary incentives (team bonuses and tournaments with prizes) to be highly effective in fostering performance. However, participants misperceive the efficacy of important elements of team governance: First, participants overestimate the efficacy of tournaments without prizes (i.e. rank / status incentives). Second, they substantially underestimate the important role of leadership for team performance. Participants' socio-economic background characteristics and participants' leadership and teamwork experiences do not fully mitigate the observed misperceptions.