Swiss programme favours mainstream work, but unconventional applicants aren’t deterred, say Charles Ayoubi and colleagues
In a 2012 Nature article, provocatively titled ‘Conform and be funded’, Joshua Nicholson and John Ioannidis showed that few of the most highly cited US biomedical scientists received funding from the country’s National Institutes of Heath. They attributed this to a reluctance at the agency to support potentially groundbreaking work. Since then, the sense that research funding is risk-averse and biased against novel work has become increasingly widespread within the scientific community.
The possible causes for this include: increasing competition for funding, making funders reluctant to gamble on unusual, high-risk studies; a growing focus among policymakers and agencies on more applied problems with obvious societal relevance over speculative blue-skies work; and the biases of expert reviewers who prefer ideas in the mainstream of their disciplines to more left-field ideas.
What has been missing, however, is concrete evidence of the existence and size of potential biases in the funding system against researchers pursuing novel ideas. In a recent study, we provide this evidence, showing that, in a prestigious Swiss funding programme, researchers with a history of publishing novel research are less likely to win grants.
The programme in question is the Sinergia grants scheme for collaborative, groundbreaking work. The scheme is aimed at large projects with an average grant of SFr1.6 million (€1.5m), and accounts for around 10 per cent of the research funding provided by the Swiss National Science Foundation (SNSF). We analysed 255 applications featuring 775 applicants, submitted between 2008 and 2012, of which 114 were funded.
We classified a scientist as novel if they had published at least one novel article before their application. To measure an article’s novelty, we searched in its reference list for combinations of cited journals that had never appeared together in a scholarly article before and had reappeared at least 10 times in other papers. This latter condition helps separate combinations of references that are simply random from those that have confirmed their usefulness in subsequent scientific works.
Previous research on defining and gauging scientific novelty shows that this type of work tends to be highly cited in the long run—suggesting high-impact research—but also tends to be published in lower-profile journals and takes longer to gather citations than traditional papers.
Comparing applicants with funding decisions, we found that teams led by researchers with a history of novel work were significantly less likely to win funding. Review panels for the Swiss scheme marked proposals on a scale of one to six: novel lead applicants scored on average 0.7 less than more conventional colleagues and were 31 per cent less likely to be funded.
This is a strikingly large effect. An important nuance is that it is only seen in teams with a high proportion of novel researchers.
We found that if less than a third of a team’s members were classed as novel researchers, it did not suffer any disadvantage with reviewers, while those with a larger share fared less well. It might be that reviewers have in mind a particular sweet spot for the make-up of research collaborations, seeing proposals from teams with a certain amount of novelty as promising, while questioning the credibility and feasibility of those from teams with a large share of novel applicants.
How does this bias affect researchers’ behaviour? One possibility is that novel researchers perceive their disadvantage and are deterred from applying to Sinergia in the first place.
But, interestingly, we found that the lower success rate of novel researchers did not discourage them from applying. In fact, we found that scientists with a novel profile were about 15 per cent more likely to apply for the grant, relative to the entire body of eligible Swiss scientists.
We have shown that researchers with a history of novel work are less likely to win funding. What we have not shown are the wider consequences, either for individual researchers or the system as a whole, or whether grant panels are justified in their decisions. Even so, in more recent calls, the SNSF has made its requirements for novelty in the Sinergia scheme more explicit to both applicants and reviewers, to try to overcome this bias.
One possibility for making funding decisions less risk-averse, which has been implemented by several funders worldwide, is to introduce a lottery element, choosing among eligible proposals at random. Another is to increase the disciplinary diversity of review panels; experts in a ‘home’ field are good at spotting a project’s flaws, but those from other fields can spot benefits that they might overlook.
Charles Ayoubi is a research fellow at the Laboratory for Innovation Science at Harvard in Massachusetts, US; Michele Pezzoni is associate professor at the Université Côte d’Azur in Nice, France; and Fabiana Visentin is assistant professor, UNU-MERIT, at Maastricht University in Maastricht, the Netherlands
This article also appeared in Research Europe