Research funding: big vs. little science

With the delicate economical situation that many developed countries are experiencing in the last years, a significant number of questions and concerns have been risen about how to properly assign and distribute funding to scientific institutions and research group leaders. In particular, a relevant question for science funding could be how to optimize the scientific “output” for a given funding level. Such a scientific output can be, of course, defined in many ways. Examples include, but are not limited to: number of patents, peer-reviewed articles, citations, Masters and Ph.D. theses defended, start-up companies based on scientific advances, spin-off technologies developed, or even groundbreaking discoveries which eventually lead to a U-turn in the mainstream view within a scientific field, and the corresponding Nobel Prize.

To obtain funding, scientific proposals have to be evaluated thoughtfully. Funding agencies, which are responsible for such a task, have to deal with tons of funding applications every year, either at the level of individuals or from big groups constituted by many scientists. An important assumption to keep in mind was that generous funding for a research group typically leads to better scientific results (due to the capacity of hiring more junior researchers, buying cutting-edge equipment, attending to conferences to interact with peers, etc). Therefore, the better the scientific project proposed, the more funding it should get and, hopefully, the better scientific output is obtained out of it.

This method does not detail, however, how should we distribute the resources to optimize the scientific output: is it better to give small grants to a lot of good projects, or, instead, to give a few large grants to a small set of exceptional high-quality projects? Different governments and agencies have their own preference. Federal agencies in many countries, such as Canada or Spain, have been using preferentially the “many small” strategy, which implies that many research groups will get funding, but each grant will be typically small. The Natural Sciences and Engineering Research Council of Canada (NSERC), for instance, funds most scientists in Canadian universities (around 62% in 2012, ¹). On the other hand, large USA funding agencies, such as the National Science Foundation (NSF), prefer the second option (the “few big” strategy), funding only the most exceptional proposals (around the 23% in 2010, ²) with much larger grants. The traditional success of universities and research centers in the USA in terms of scientific impact naturally leads to think that the “few big” strategy was more appropriate, since it promotes competition between researchers to surpass their colleagues, and the best researchers receive lots of funding to develop their promising projects. However, this idea was not tested until recently.

In a work published this year in the journal PLoS One ³, Jean-Michel Fortin and David Currie, researchers from the University of Ottawa in Canada, examined the scientific impact of individual Canadian researchers over a period of four years, and related their performance with the level of funding received (via NSERC grants). Different indicators (number of article published, number of citations received, the most cited article, and the number of highly cited articles) were used to determine the impact of each researcher. The scientific impact determined with this method was, therefore, indicative of the influence of each researcher in the scientific community (and therefore resembling his/her contribution in the collective effort of the scientific community).

The results pointed out that, as expected, scientific productivity increases with funding. But surprisingly, the increase was notably weak, meaning that the productivity of a group is not doubled when its funding gets doubled. In other words, the global scientific output obtained when we give, for example, a grant of 100.000 to five good researchers.

The researchers also considered groups which received extra funding from other agencies (such as the Canadian Institutes for Health Research, CIHR), and they found that these groups were not, on average, more productive than the rest of the groups. Impact was, in all cases, a decelerating function of funding. Or, in other words, the impact-per-dollar was lower for large-grant holders.

The conclusions of this study are of vital importance for the design of optimal funding strategies for scientific research. In particular, these results do not support the traditional hypothesis that “larger grants lead to larger discoveries”. On the contrary, funding strategies that target diversity, in the form of many small grants awarded to a large number of research groups, are likely to be more productive than the classical approach of “targeting the excellence” of a few groups.

As a personal note, I consider that this take-home message might be especially important for countries in difficult economical situations and which plan to reform their science funding strategies. For instance, in a recent note to Nature, Carmen Vela, the Spanish science secretary, stated that they “will reduce the number of grants offered each year in the Ramón y Cajal tenure-track programme […] However, the quality of each grant will improve”. The results presented by Fortin and Currie suggests that such an strategy might not be convenient if Spain wants to improve, or even maintain, its level of scientific productivity. They also show that, especially when publishing in a scientific journal like Nature, even politicians should ground their claims on solid data.