Large Teams Develop and Disrupt Technologies
30 Pages Posted: 4 Jun 2019
Date Written: May 15, 2019
Although individuals used to be the main sources of scientific and technological breakthroughs, team collaboration has become a “universal trend” in science and technology. This raises an important question of whether larger teams generate “better” knowledge than smaller ones. Recently, Wu, Wang, and Evans (hereafter “WWE”) built on Funk and Owen-Smith’s CD index to capture scientific and technological works’ contribution besides citation, and reported to have identified a nearly universal pattern that small teams disrupt and large teams develop science and technology across disciplines. This study has garnered broad attention and is likely to have strong implications in research, management, and policy. We notice that the WWE study contradicts several published works, which in fact show that large teams are associated with novel works that disrupt existing science and technology. Such discrepancy is likely caused by two limitations in the WWE study: (1) it focuses on the net effect of how a technology disrupts and develops its knowledge base, implying that as a technology becomes more disruptive, it gets less developmental, and (2) it considers all prior arts of a technology as a whole. As a result, technologies that are both disruptive and developmental (i.e., dualistic) cannot be identified. Conclusions drawn from this misclassification can be misleading and cause inappropriate allocation of critical human and capital resources. In this commentary, we argue that a technology can disrupt a part of its knowledge base and develop another part simultaneously, decompose the WWE index into a Disruption index and a Development index for measuring a technology’s contribution and re-analyze the same patent dataset used in the WWE study. Our results show that compared with small teams, large teams are more likely to develop and disrupt previous technologies.
Keywords: Disruption Index, Development Index, Team Size
Suggested Citation: Suggested Citation