Collaboration in crowdsourcing contests: how different levels of collaboration affect team performance.

With the advances in internet technologies and the emergence of crowdsourcing, organisations are now increasingly looking outside their boundaries for solving problems. Yet, the success of crowdsourcing depends on high-quality participation of crowdsourcing individuals and teams. In recent studies, examining the effect of simultaneous collaboration and competition mechanisms on individual and team performance in crowdsourcing has received considerable attention. But, none of these studies examined how different levels of collaboration affect team performance in crowdsourcing contests. In this paper, using a rich data set from a crowdsourcing platform, Kaggle.com, we study how team discussion-forum performance and solution-sharing performance affect its performance in online crowdsourcing contests. Our results suggest that team's discussion-forum performance and solution-sharing performance have significant effect on its competition performance. Our findings offer valuable theoretical and managerial implications for researchers and crowdsourcing sponsors who want to improve teams' performance in crowdsourcing contests. [ABSTRACT FROM AUTHOR]