Over time, biomedical scientists tend to choose less risky research topics, concludes a study published 9 November in the Proceedings of the National Academy of Sciences (PNAS). An examination of millions of studies of molecules that play a role in biological function reveals that “scientists in this field were about 10-fold less efficient [at making new discoveries] than they could have been,” one of the paper’s co-authors, computational biologist Andrey Rzhetsky of the University of Chicago (UC) in Illinois, told our colleague John Bohannon on publication day.
That efficiency penalty is “compared to a scenario where every scientist followed the optimum risk strategy, and science had full disclosure, where the results of every experiment were published, including all negative results,” added co-author James Evans, a UC sociologist, in the interview. Therefore, to increase the efficiency of research, Rzhetsky continued, “[s]cientists should take on more risk in general. The goal is to do work that gets cited, so you have to make discoveries.”
The need to win grants and tenure encourages researchers to pursue conservative projects that are likely to produce positive results and publications.
The incentives built into the funding and tenure systems, however, militate against such bold action, the authors recognize. As other experts have also repeatedly warned, the need to win grants and tenure encourages researchers to pursue conservative projects that are likely to produce positive results and publications, rather than the risky ones that may result either in striking scientific advances—or in unsuccessful outcomes that journals are unlikely to publish.
Some scientists do, however, undertake the bold experiments that produce notable breakthroughs—and occasionally a Nobel Prize or another prestigious award. In a companion study published online on 1 September in the American Sociological Review, Evans, Rzhetsky, and Jacob Foster—an assistant professor of sociology at the University of California, Los Angeles, who is also a co-author of the PNAS paper—found that this possibility of prize-winning is “the most plausible explanation” for why researchers take the risks they do.
Research that breaks out of established boundaries is, of course, “more likely to achieve high impact and recognition,” including, for example, major honors such as the Nobel Prize, the authors note in the paper—but it is also “more likely to be ignored.” So, given the current system of incentives, “the additional reward is insufficient to compensate for the additional risk,” they write. Specifically, the authors found that researchers were six times likelier to perform experiments that covered known ground than to strike out in unexplored directions, and that this proportion remained steady even as scientific opportunities grew rapidly over the quarter century covered by the study. A major “reason why unexpected findings that change the landscape of science are so infrequent [is because] pursuing them is a gamble, without enough citation payoff, on average, to justify the risk” of wasted time and lost opportunities for funding and advancement, the authors write.
The authors do not argue that “risky” research is the only science that is worth doing. “‘Normal’ science that characterizes a known relationship more deeply has its own value,” they write. Given the tremendous importance of innovative work for the advancement of science, however, they suggest “two policy levers to promote risk-taking”—which have been suggested by others as well. “Since career pressure is a powerful incentive for conservative behavior,” the authors write, “decoupling job security from productivity early on can encourage originality. … So can funding scientists rather than projects. … The other lever sits outside the tension between productivity and posterity: funding agencies can lower the barriers to risky projects by funding them more aggressively.” It’s possible, they believe, that “these interventions may be able to counter the stable conservatism” that governs much of the science being conducted today and free researchers from the constraints on risk-taking that results from the current structure of incentives.