مکانیسم های همکاری در تیم های مبتکر : ترکیب، شبکه های اجتماعی، و جغرافیا

This paper investigates the composition of creative teams of academic scientists engaged in inventive activity. Our data provides a unique opportunity to explore the links between team composition and commercialization outcomes. We find that there are coordination costs associated with reaching across academic departments and organizational boundaries to build teams. However, we also find evidence of benefits due to knowledge diversity, particularly in the cases of truly novel combinations. In support of internal cohesion arguments, we find that performance improves with the experience of the team. In line with arguments regarding the value of diverse external networks, we find that teams that are composed of members from multiple institutions – focal university, other research institution, and/or industry – are more successful in generating patents, licenses, and royalties. Finally, we find that the presence of prior social ties supporting links with external team members positively influences commercial outcomes. We find that there is no benefit to proximity in team configuration.

Invention, in spite of the romantic image of lone genius, has increasingly become a team endeavor. Problem-focused creative teams involving individuals with varied backgrounds are a staple across organizations, including academic institutions, small entrepreneurial ventures, and large corporations (Reagans and Zuckerman, 2001 and Roberts, 1991). Creative teams have become especially important in research and development, inventive efforts, and new product development as scientific activity is becoming more specialized (Wuchty et al., 2007). Creating valuable and novel solutions requires melding multiple types of individual expertise. One notable fact is that team size among American inventors, as witnessed by the number of inventors on U.S. patents, has been increasing at the rate of 17% per decade (Jones, 2009). Technical innovation is increasingly at the intersection of traditional domains of knowledge calling for greater use of interdisciplinary creative teams. Simultaneously with the growth in team size, there is also a trend towards including individuals from outside the focal organization in order to tap external expertise (Chesbrough, 2003). Despite the pervasiveness and importance of teams, many open questions remain as to how to successfully configure effective teams. Issues of team configuration become even more salient when the task is complex and requires creativity and problem solving (Amibile, 1988). The desired outcome for commercially oriented R&D teams is the generation of an invention that is novel, valuable and non-obvious. While organizations have an interest in finding team configurations that increase the probability that scientific and economic value result, the relationships between combinations of individual expertise, expertise diversity and team performance have proven difficult to disentangle (Williams and O’Reilly, 1998). In addition, greater understanding of the social networks that underlie these combinations is needed as team learning capacity, and hence team performance, may be influenced by these social ties (Reagans et al., 2005). One vexing problem in evaluating team configuration is lack of systematic data on team performance outcomes. The objective of this paper is to enhance our understanding of the links between team structure and outcomes. Our subject is academic teams of university scientists and other external members who engage in inventive activity. In the context we study, team composition is internally managed rather than externally assigned or determined. This means that teams are able to self-organize, providing an ability to experiment with different configurations of individuals, including adding members from different departments or even other organizations. We have detailed data on the individuals that comprise the invention team. We are able to follow inventive teams from the initial reporting or disclosure of their invention and the progress the idea makes towards realizing commercial value. In this process there are a variety of outcome measures such as the granting of a patent based on the invention, the subsequent licensing of that intellectual property to a commercial firm, and finally, the generation of royalties from the license. Thus, we can test how different constructs of team composition affect team performance using econometric methods. This provides a unique opportunity to study how team composition affects outcomes and productivity. The paper is organized as follows. Section 2 provides a review of the literature and develops hypotheses about creative team composition and effectiveness. Section 3 introduces our data and study context and develops our empirical measures. Section 4 presents results and Section 5 concludes.

Scientists, particularly those in the medical and life sciences fields, rarely work alone (Wuchty et al., 2007). Rather, research activities are generally conducted in teams that bring together numerous individuals across knowledge domains and organizational boundaries. Teams, however, are not homogenous, differing on many dimensions including team size, team experience, knowledge combination characteristics, organizational diversity, as well as social grounding. An open and important question is, what factors differentiate more successful innovation teams from less successful innovation teams? In this paper, we probe the composition of academic research teams, developing and testing a set of theoretically derived hypotheses that link team attributes to performance. Our results provide insights into how to configure creative teams to increase the likelihood of a successful outcome. We first explore two knowledge-based considerations argued to influence team composition choices – recombination potential and ability to span scientific and innovation logics. With respect to the first factor, we find that there are benefits associated with knowledge combination novelty. Exploration, a broad search that leads to the integration of knowledge in rarely tried combinations, offers a higher probability of generating significant breakthroughs. The likelihood that an invention is patented and licensed increased with the degree of novelty in the combination of knowledge expertise on the team. While the results for licensing income were not statistically significant, this may be due to the fact that truly novel inventions require more time to realize their commercial potential (Rosenberg, 1974). The data also show that there are gains to including individuals from external organizations on the team, particularly if these links offer access to market-related as well as science-related knowledge. Specifically, having a team member from industry significantly increases the probability that a disclosure will have a positive outcome with regards to patenting and licensing, as well as having a positive relationship with the level of royalties generated. The value in linking to an industry-based external network seems to be one of identifying commercial opportunities and accessing financial resources, as well as extending scientific knowledge. The benefits of ties with external academics appear to be narrower – contributing mainly to the latter (scientific advances), but not necessarily the former. Though technological and knowledge-related issues may catalyze team formation and shape team composition, understanding coordination issues and the social relationships associated with these combinations are equally important. Progress towards a goal in a team environment requires the coordination of effort. We find that the challenge of achieving such coordination increases, and the likelihood of innovative success decreases, the greater number of departmental and organizational boundaries spanned by the team. Two socially based attributes – experience and embeddedness – provide means to meld a set of diverse individuals into a coherent and productive team. The data suggest that through repeated interactions, teams develop coordination capabilities, communication mechanisms, and task routines that enhance commercial performance. Relatedly, an external team member's initial willingness to cooperate in the transfer of knowledge appears to be a function of the level of attachment, or embeddednesss, in his ties to others on the team. The data shows that for both external academic scientists and external industry scientists, having a pre-existing social tie increases the innovation performance of the team. Perhaps most interesting, we find that the type of pre-existing tie is also material. Further, the existence of a prior social tie provides a foundation for collaboration across distances. The academic teams that we study provide a transparent context to investigate the relationship between the team composition and team innovative performance. The teams we observe are self-organizing as there is no administrative assignment of individuals to participate in inventive activity. Team membership on an academic invention disclosure is dictated by the contribution of the individual to the effort and all individuals listed as a member of the inventive team will have made material contributions. The examination of teams that are more organic in organization and more selective in their membership allows a better comparison of how different configurations of individuals perform. We believe, however, that the results we find are generalizable to other settings. Though academic inventive teams have more freedom in choosing with whom they collaborate, the knowledge-based factors that drive composition choices along with the coordination challenges associated with these choices are similar to those faced by other research-intensive teams in private companies and government labs. In particular, many firms that conduct research in emerging fields find themselves operating within Pasteur's quadrant and dealing with significant pressures to integrate science and market logics (Gittelman and Kogut, 2003). In such environments, the ability to access knowledge from external players and to integrate knowledge from multiple disciplines in novel ways is valued as it has been shown that such capabilities are positively associated with research outcomes and competitive performance (Henderson and Cockburn, 1994, Rosenkopf and Nerkar, 2001 and Chesbrough, 2003). Less hierarchical organizations tend to be similarly fluid to universities in the organization of inventive activity. More hierarchical organizations who prescribe how this activity will be organized may benefit by adding domain novelty and by encouraging repeat interactions. Further work will explore in more depth the different types of team configurations. Our results suggest that academic scientists learn how to organize and work with teams that are more likely to generate commercial success. We expect that inventive teams follow more or less archetypical patterns as it appears that some faculty members like to work with subordinates while other faculty members prefer to work with their peers. The patterns suggest that some inventors are more or less monogamous and work with the same individuals while other scientists reconfigure their teams more frequently. It may be that these patterns affect performance or that certain individuals are most productive once they find their type of team. In future work, we hope to examine how team structure evolves over time in order to increase the probability of a positive outcome. Innovation is increasingly becoming a team sport. And like all team sports, success is a function of the expertise of the individual players, a solid roster enabling coverage of the key positions with the potential of a few stellar combinations, and an integrating set of social ties that enables the individuals to function smoothly as a unit.