The science behind team science

"It turns out there's a lot of similarity between the collaborative structures needed to put on a Broadway show and to do team science," Bonnie Spring, PhD, told participants at APA's 2011 Education Leadership Conference.

The average number of people involved in creating a Broadway show increased from two around 1880 to eight around 1930 and has stayed steady since then, she explained, citing a 2005 study in Science. Eight seems to be "something of a ceiling" when it comes to teams, said Spring, adding that the number of authors of scientific publications has jumped from one to three or four over the last four decades. The Broadway literature also suggests there's a "sweet spot" between having too many old-timers or newcomers on teams, she said.

While Broadway's experience may foreshadow what's to come in research, Spring said, the emerging science of team science has already produced important lessons.

When it comes to collaborations among universities, for example, research points to the need for more effort put into bringing collaborators closer together, if only virtually.

Publications jointly produced by collaborations of two or more different institutions tend to have more impact than those produced by single schools or solo authors, said Spring. But a review of National Science Foundation-funded collaborations found that the more universities involved, the fewer the publications, patents and grant awards that result.

That's because a collaboration's effectiveness declines when team members are more than 30 meters apart, Spring explained.

"There are no water cooler conversations," she said. "If you want them to have the conversations and collaboration that drive team science, you have to make up for that lack of bumping into each other."

Strategies for overcoming that hurdle include using Basecamp and similar project management tools, exchanging students across labs and holding in-person meetings, said Spring. Regular interactions focused on the science itself also facilitate team-based research, she added.

Teams also need to develop a common language and constructs, something that's harder than it sounds, Spring warned. Other disciplines may have methodologies and assumptions very different from psychology's, she said.

Institutional cultural norms that support interdisciplinary teams are also key. In addition to institutional leadership, tenure and promotion committees need to understand the differences between interdisciplinary and solo work. In psychology journals, the first author is traditionally the student and the second the mentor, Spring pointed out, while in medical journals, it's the first and last authors who matter. "Those are small things that make a big difference," she said.

To capture all these lessons, Spring developed a series of online training modules aimed at junior investigators seeking their first grants, senior investigators developing collaborative projects and research development officers. Funded in part by the National Institutes of Health, the free interactive modules include an overview of the science of team science plus modules focused on the team research process in behavioral science, basic medical science and clinical medical science, available at teamscience.net.

—R.A. Clay