Safe Bets and Risky Propositions (NSF)

Funded by the National Science Foundation.
Award number SMA- 1856090

Leveraging Rich Data to Understand Scientific Diversity, Impact, and Potential of Teams

Innovation in science and technology is often the province of teams. As knowledge becomes ever more specialized, teams can tackle complex problems whose solutions require insight from multiple domains. However, teams are fundamentally social entities. Gaining the benefits of diverse teams for solving hard problems requires bridging knowledge that resides in the minds of multiple individuals. Research shows that while promising in theory, in practice, diverse teams often fail to realize their potential. Diversity poses a puzzle to team science. Given the historical evidence supporting the value of interdisciplinarity, how can we distinguish the combinations and overlap of expertise possessed by individuals and by teams who are less productive, from those whose work will go on to fuel decades of important discoveries? That is the question we seek to answer.

This project has intellectual merit along three dimensions. First and foremost, insights from this project come at a critical point in time. Society is on the cusp of major breakthroughs in many fields (e.g., brain science, artificial intelligence). Understanding how to find the safe(r) bets from among the “high-risk high reward” big ideas represents a major scientific advance needed to advance the scientific enterprise itself. Second, we develop “full text” based methods for placing scientific expertise and identifying processual markers to quantify ideational co-creations. Existing work relies on structural markers, and has yet to peer into the ideas themselves. Applying recent developments in machine learning, such as Doc2Vec, the full text of publications and patents is used to operationalize expertise and diversity in ways that otherwise would not be possible. Third, we advance basic knowledge of diversity in teams. The studies performed in this research program represent the largest empirical studies conducted to date on team diversity, process, and performance. Whereas the science of science will advance from having a stronger evidentiary basis for predicting team success, so too will the science of teams, which provides the foundation for recommendations for forming teams in healthcare, disaster response, military, and workplaces.

Additionally, this project has many dimensions of broader impact:

1) the project will enable society to more efficiently allocate scientific expenditures, making safer bets on risky science. This research will provide guidance to decision-makers on balancing investments in training and funding interdisciplinary scholars or teams, gauging the likely success of a team submitting a proposal, evaluating the progress of a team once it has started to produce papers thereby informing decisions on renewals and additional funding.

2) the project can benefit the careers of scientists, helping them to make informed decisions on forming teams that will enable them to have more productive careers.

3) the project answers questions about how the contributions of minorities, females, and junior scholars who work on science teams are grown into the ideational products, the papers and patents of those teams. This allows us to identify the situations where participation on a team is more and less beneficial to the minority/female/junior member.

4) we will make derived and anonymized data available through data use agreements.

5) all of the metrics developed and refined will be disseminated in publications.

6) we develop as part of the project, two online modules, one for female faculty and one for junior faculty members. The content will be based on current research and insights from this project and provide concrete recommendations on how and when collaboration affects the impact of these groups.

Whalen, R., Lungeanu, A., DeChurch, L. A., & Contractor N. (in press). Patent Similarity data and innovation metrics. Journal of Empirical Legal Studies.