Want disruptive research? Go small instead of big

Smaller research teams conduct more disruptive research; a new study could change research funding allocations. Shutterstock

Want disruptive research? Go small instead of big

Conventional wisdom suggests that large teams are better at solving complex problems. And, in this era of specialization, many discoveries are the work of large groups of experts from a diversity of fields. But it’s not yet time to abandon support for small teams of scientists, according to a new study in Nature.

By analyzing data on the work of more than 50 million teams in science and technology, researchers discovered that larger teams developed recent, popular ideas, while small teams disrupted the system by drawing on older and less prevalent ideas.

Disruptive research tends to introduce new approaches and ask fundamental questions, while “developmental” research is more likely to adjust or test old theories and apply them in new contexts.

1,000 authors

Examples of very large projects include the Human Genome Project, and more recently, the projects to detect gravitational waves and the Higgs boson particle.

The detection of gravitational waves — a discovery that was published in a paper with more than 1,000 authors and which received the 2017 Nobel Prize in physics — “could possibly have been the most conservative experiment in history,” says James Evans, the senior author of the new Nature paper and a sociologist at the University of Chicago. “It tested a 100-year old hypothesis and that hypothesis was generated by one person, Albert Einstein.”

Tech Insider: Scientists won the Nobel prize for detecting gravitational waves.

Solo scientists like Einstein, or small teams, appear to come up with novel ideas that change the course of a field. Those are becoming rarer, though: authorship lists on scientific papers have grown in the last century, from about one author per paper in 1913 to an average of 5.4 authors per paper in 2013.

The impact of this shift in team sizes isn’t completely known. This new study documents the different roles that small and large scientific teams play in the research landscape, but it raises more questions than it answers.

Different sizes, different approaches

These are questions that would help funding agencies to make better decisions.

“There’s this long, long debate about this,” says John Walsh, who studies science, technology and innovation using a sociological perspective. “Is giving a lot of money to one (large) project a good way of moving the science forward, or is it better to give lots of people more modest funds and have them work on different things?”

In small teams, people are more apt to take chances because the cost of taking a chance is lower. There are fewer monetary resources invested and fewer careers at stake, Walsh explains.

What’s interesting, adds Walsh, is that the effects described by Evans and his co-authors Lingfei Wu and Dashun Wang show up even at modest team sizes — between one and 10 people.

One way of interpreting the finding is that small teams have a better chance of finding something unusual because they can be nimble and adapt to new findings by changing direction and pursuing new paths as they open up. It’s unclear, though, whether small teams propose more innovative and disruptive ideas to begin with, or whether they are more likely to change course midstream and “benefit from serendipity,” he says.

In contrast, large teams are more like huge shipping barges — impossible to turn on a dime. They are also faced with all sorts of conservative pressures.

“We realized just from our own experience that creating these big federations of people ends up really stifling certain kinds of ideas and certainly stifling the likelihood of following an interesting or unusual path,” says Evans. You have to get to the common denominator to build consensus, and “the common denominator, when you have a lot of people, is yesterday’s hits.”

Indeed, Evans’s study shows that large teams are more likely to cite the really famous older papers, whereas smaller groups are likely to cite a broader array of papers and to resurrect some more obscure findings from prior literature.

Because of the burden of co-ordination in large teams, “it’s much more likely that a small group of committed people can hammer on a problem and come up with a breakthrough or disruptive solution than a really large group, where they’re not going to be able to to really coordinate,” says Steve Kozlowski, a professor of organizational psychology.

The co-ordination challenges increase when the large group is interdisciplinary because scientists have different sets of assumptions about the way the world works based on their disciplinary training.

Nature podcast: Social scientists James Evans and Pierre Azoulay discuss how team size can affect research outputs.

Size matters?

Mirta Galesic, professor of Human Social Dynamics at the Santa Fe Institute, thinks that the small and large teams may represent different stages of the natural history of an idea.

Initially, a disruptive or unconventional idea is born small and only has a few people working on it. But if it stands up to initial investigation and scrutiny, it may attract more funding and more scientists to work on it. In other words, the disruptive work of small teams represents the seeds from which big projects grow.

“I think it’s possible that the small and large teams occur at different stages of the scientific process and that it could be a case that the size is correlated with the process, rather than the cause of a disruption,” says Galesic.

Implications for funding agencies

“What’s the secret sauce that the small teams seem to have?” Kozlowski asks. He’d like to see funding agencies invest more resources in studying team science: “If we’re going to be pushing for these large investments to tackle big problems, then we want to have research to help inform how these larger teams should be set up and managed.”

Research teams of varying sizes should be funded, a new study suggests. Shutterstock

Some agencies are still supporting smaller teams. For example, studies by the National Institute of General Medical Sciences (NIGMS) a few years ago showed diminishing marginal returns on grant funding. The NIGMS researchers argued that this provided a rationale for spreading the research money across more labs.

Given the distinct roles small and large teams play in moving science forward, “our findings suggest the importance of supporting both small and large teams for the sustainable vitality of science and technology,” the authors write in the paper.

In particular, small teams are often neglected, and so the study is “an encouragement for funders to realize that if they want to fund disruptive innovations, they’re going to have to take more risks, and smaller teams are one important dimension of risk that we’re suggesting they should consider,” Evans says.

The current emphasis on funding large teams, he says, has a long term unintended consequence: “science as a whole ends up looking more conservative today, and I think it’s starving future innovations.”