Abstract
There is too much information available to ever know everything about a subject, let alone assimilate the evidence into something meaningful.
This was the concern among intellectuals in the mid-15th century as a result of the impact of Gutenberg’s printing press. 1 While this fear has lingered for centuries, the age of the internet and 24-h connectivity makes all before pale into insignificance. It is estimated that 90% of the world’s data has been produced in the last two years and this process is likely to continue. The expansion of information has parallels in medical literature, with an increasing number of journals and a reduction in the cost of publication. At the time of writing, PubMed had over 24 million articles catalogued. This is fuelled by the incentive to publish the research community places on academics and clinicians, limiting their worth to the number of publications per year, thereby encouraging quantity over quality.
The broadening of medical research is generally a good thing but it is becoming increasingly difficult to keep abreast of a subject without becoming super-specialised. Some suggest that the advent of new technologies, such as via social media, can help to navigate this, but increasingly keeping up to date in a subject is more of a chance process. This places greater importance on reviews/systematic reviews, but these are becoming more costly and complicated to produce. Despite the benefits of large numbers of publications, being able to adequately read, critique and synthesise research is technically difficult and time-consuming. Knowing a subject in its entirety may never be possible.
As a solution to this, we speculate on a proposal in which a limit is imposed on the number of papers a person can publish in a year. The number would be relatively small and may vary from one subject area to another. Publishing of trial protocols, some methodology papers and safety information would be exempt from this limit. We do not propose a reduction in research funding nor only publication of positive results (this can be enforced by the journal editors). The goal of the proposal is to move from quantity to quality by removing the incentive for researchers to have vast publication lists. Of course some people achieve the ‘quality and quantity’ benchmark but all too often one is sacrificed for the other. This hypothetical scenario can help to point out the flaws in the current model.
Why would a reduced number of publications be useful? For the reader, the overall number of publications in an area is likely to fall, and probably present in less outlets, making it easier to find relevant and important material. It would help the non-specialised understand a subject, rather than being misled by small, inconclusive research outputs. This would happen in different ways. It would stop the artificial division of research into numerous segments; it would reduce repeat publications of data in similar papers; and it would potentially preclude smaller studies from being published. The last of these factors may reduce the overall evidence available, but this could be beneficial as smaller studies often have spurious results and would be an important factor in the planning stage of research as a disincentive to embark on underpowered studies and produce unspecified post hoc analyses.
For the researcher, the currency of research is publication numbers (combined with the enigmatic impact factor) and an academic’s career progression depends on this. Lessening the importance of authorship would act to increase the importance of other parameters. Instead of the number of publications, the creation of datasets, study quality, the completion of reports, sharing of data, replication by others and acknowledgements become currencies of success. 2 Authorship becomes one marker of achievement among the many other aspects of research. By removing this incentive, this would allow researchers to focus on the studies they do and encourage them to take greater care over what they produce. The ground-breaking research paper is worth many insignificant ones. A reduction in numbers should not correspond to a reduction in detail. There is plenty of room for online storage of information. For the less productive members of the community, achieving a minimum number could potentially act as an incentive.
Limiting the number of publications may have negative consequences, the most important of which is holding back or not publishing important findings. People may try to game the system and delay publications to other more lean years to fill their ‘quota’, but important findings should still be published. It would also be very difficult to police. Restriction of publications would not prevent the problem of people producing a small number of poor quality research articles, but changing the emphasis to quality should act as a motive. Limiting the number of publications would reduce the long lists of authors on some publications. This could reduce the number of non-contributing authors but also narrow the expertise leading to poorer quality. Senior authors may avoid the less famous journals leading to increasing entrenchment of some over others. For the journals, limiting publications may result in the death of some, but the fittest (or probably the most famous) would survive. So long as enough diversity exists to avoid a monopoly, this may not be detrimental.
Our proposal to limit the number of publications is more of a thought experiment and would need strict safeguards to ensure important outputs are not restricted. However, it does highlight important problems with ever greater publication and we urge readers to rethink the incentives in academia. Our suggestion goes against the trend to publish everything that many advocate, but quality and care really do count more than a numbers game. As Douglas Altman famously wrote, ‘we need less research, better research, and research done for the right reasons’. 3 We urge a shift away from having to publish a large number of articles, to the publishing of high quality research.