If you leave a plant in a dark closet and it dies, you can probably guess what happened. But to investigate it scientifically, you would need an identical plant, with the same soil and water, placed in a bright window for comparison.
Scientists call this a controlled experiment. It’s the gold standard for research, although it usually requires a bit more than two plants and a closet.
But some things can’t be neatly analyzed in a lab.
“It is difficult to experimentally manipulate what goes on in the real economy. And so you’re left with data that just appears naturally,” says Kevin Boudreau, an associate professor of entrepreneurship and innovation at the D’Amore-McKim School of Business and a research associate at the National Bureau of Economic Research. “And nature is an awful experimentalist.”
Boudreau, who is also affiliated with the College of Computer and Information Science, is at the forefront of a generation of scientists who use the scientific method, with randomized, controlled trials, to get at the root of why scientists, innovators, and entrepreneurs behave in certain ways. Their work can provide insights to improve the process of innovation and science in the future.
He researches the researchers, if you will.
Boudreau began his career as an engineer, running experiments in microelectronics. When he moved into the social sciences, he brought his research skills with him. Boudreau’s more recent work, featured in an article in Science, has included studies on unintentional bias in the peer review process and investigations into how scientists choose who to collaborate with.
But experimenting on scientists has its challenges.
“It’s really hard to imagine trying to treat scientists like lab rats,” Boudreau said.
Because he is studying scientists and innovators in the process of doing their work, Boudreau has to be especially careful that his experiments don’t interfere with his subjects.
“You’ve got to make sure there’s a basic fairness in what you’re doing,” Boudreau said. An experiment can’t hamper one group of scientists while helping another. And at the end of it, Boudreau always intends his research to have a positive effect on whatever organization has agreed to participate.
“We generally leave our industry and governmental partners saving money, making new discoveries, and adding to their bottom line,” he said.
In one such experiment, Boudreau worked with administrators to adjust the requirements for a university grant to include attendance at one of several information sessions. Partway through the session, attendees were randomly split into smaller groups to discuss their research interests. When the grant applications came in four weeks later, Boudreau checked which researchers had teamed up.
Boudreau found that researchers, despite having access to a university database of potential collaborators, were 75 percent more likely to form teams with people they met face-to-face in the small groups.
“Even in this environment, where people have the best information possible to find new collaborators, they simply don’t,” Boudreau said. “There are limits to what a database can convey. Having a real human interaction has a dramatically greater effect.”
Boudreau is currently collaborating with researchers around the world on several large-scale field studies, including a study to determine which types of researchers are more likely to collaborate with people outside their fields, and which unexpected effects might occur as universities try to encourage this type of research.
By applying the scientific method to study the mechanisms behind science and innovation, Boudreau says he can provide universities and other organizations with the opportunity to create policies that better serve their researchers and their work.
“Here at Northeastern, we have this opportunity to rethink how social science, economics, and business get done,” Boudreau said. “Applying experiments in the field is an incredibly powerful way to do things.”