Source: James Fryer
You’re applying for an academic job in the UK. What factors do you think are most likely to decide whether or not you’ll get it? Your reputation among colleagues, your ability to teach, your collegiality, your creativity? In most research-intensive institutions two factors will trump all of the above: publications and research income. Since the job of a researcher is to do research and publish it, these may seem reasonable criteria, but they have become problematic both in terms of how they are interpreted, and in the way they have become all-important.
Looking first at publications, it has long been recognised that number of publications is not a useful way of assessing researchers. In the research excellence framework and its predecessors, individuals were allowed to submit up to four publications – a move that should encourage a focus on producing a smaller number of meaty publications rather than disgorging findings in tiny gobbets. Many funders and some employers take a similar approach: applicants for grants or jobs are asked to nominate a specified number of publications rather than all publications – thereby disadvantaging those who adopt a “never mind the quality, feel the width” approach.
So we need to evaluate quality rather than quantity, but how? Many employers say they want applicants with publications in “high-impact journals”, but this is fraught with problems. These are well-articulated in the San Francisco Declaration on Research Assessment, a document that has been signed by many august individuals and institutions, including the Higher Education Funding Council for England, the Royal Society and the British Academy. Nevertheless, the belief in journal impact factors remains entrenched in many places, leading researchers to waste months if not years in a demoralising cycle of submission to journals with ever-declining impact factors until at last a home is found. This is bad for science, and not just because it delays the publication of research. The problem with high-impact journals is that they regard newsworthiness as a major criterion for deciding what to publish. Although they aim for methodological rigour, this tends to take second place to “interest”. Negative results from well-designed studies are unlikely to find a home in a high-impact journal, and so the literature paints a distorted picture of reality. Furthermore, the highest-impact journals often have a lower standard of reviewing than the lower-impact specialist journals, because the editor doesn’t have expert knowledge of the subject area. Finally, high-impact journals usually take only very short papers; methods may be relegated to supplementary material, making it easy to highlight the most exciting results while glossing over the detail, where the devil is often lurking.
Reliance on grant income as an indicator of research prowess is even more corrosive. For many universities, a high proportion of their core funding is linked to research grants. Consequently, we now have the weird situation whereby a researcher who achieves important results with little or no funding is valued less than someone who receives a huge grant but fails to do anything sensible with it. It should be a matter of concern to funders that in contemporary academic life, people are encouraged to write expensive grant proposals rather than thrifty ones. The other trend is the rise of the “research magnate” who accumulates grants as if they were pots of gold, but then is overwhelmed by the workload of the resulting research portfolio. Some top researchers from the past would not have flourished in the current system, because their research was not expensive enough. After a career spanning 40 years, Daniel Kahneman’s elegant experiments led to a 2002 Nobel prize, but they did not require costly equipment or large squads of staff. This kind of research would be devalued in the current system for not generating enough research income.
But there is a deeper concern about changes in our scientific culture. The system of valuing high-impact publications and expensive grants has rewarded those who achieve these goals, and who have a vested interest in perpetuating the status quo. In effect, we may be driving out the very people we need to retain: those who are interested in science as an end in itself, rather than as a way of achieving personal advancement. If a scientist has the option of publishing a detailed account of a piece of work in a middle-ranking journal, or a shortened version that omits problematic findings in a high-impact journal, many feel that it would be career suicide to go for the former option. If they want to carefully analyse, reflect on and write up their current research before applying for further funding, they may find line managers threatening them with redundancy.
There is growing recognition of these problems. In April I’ll be chairing a joint meeting of the Academy of Medical Sciences, the Wellcome Trust, the Biotechnology and Biological Sciences Research Council and the Medical Research Council on research reproducibility, where the topic of incentives will be high on the agenda. As Ottoline Leyser, deputy chair of the Nuffield Council on Bioethics, has put it, the “relentless focus” on publishing in prestigious journals encourages poor practices such as “over-claiming the significance of research findings, sticking to trendy areas of science and leaving important but confirmatory results unpublished”. If our focus remains so narrow, we also risk losing sight of the purpose and meaning of science itself.
Dorothy Bishop is professor of developmental neuropsychology at the University of Oxford.