Research Says Science Funders Should Allow Failure

Print Share on LinkedIn More

June 20, 2011; Source: Boston Globe | Funders who want to see the money they funnel into scientific research pay big dividends need to get comfortable with failure. According to the Boston Globe, new research from MIT economists finds that scientists who are given the freedom to fail are motivated to undertake "the kind of exploration that ultimately leads to big breakthroughs and innovation."

For their study MIT researchers studied outcomes of two groups of biologists. One comprised of investigators receiving money from the nonprofit Howard Hughes Medical Institute, which the Globe says, makes "long-term open-ended research funding, with a philosophy that is often summed up as funding 'people, not projects.'" The other group was made up of investigators funded by the federal National Institute of Health that typically funds specific research projects that scientists propose.

The study found that those doing open-ended research produced twice as many highly influential papers, as well as "more duds." More importantly, the study suggests that given the freedom to depart from previous research or purposely choosing to delve into an area where less is known, while increasing the chance of failure, such endeavors also could result in a "big leap forward."

One scientist, Jack Szostak, a professor of genetics at Harvard Medical School and Massachusetts General Hospital, has been supported by the Howard Hughes Medical Institute since 1998. He said being able to pursue projects without restrictions is "every scientist’s dream . . . It lets me do work I think would be hard to do otherwise."

While the study suggests the benefits of letting scientists have leeway in what they research and how they approach their work, no one is suggesting that this should be the only funding model. Dr. Francis S. Collins, director of the NIH and a scientist whose past work was also supported by the Howard Hughes Medical Institute, said he believes that scientific research should be a "mix" of approaches. He advises not "to put all your resources in one model."—Bruce Trachtenberg