Select Committee on Science and Technology Written Evidence


Memorandum from Nature

  Nature receives over 10,000 submitted papers per year, of which we publish about 1,000. Since we were launched in 1869, we have had no editorial board of independent scientists, as many journals do; we make our editorial decisions on the basis of our own scientific judgment.

  We reject about 60% of submissions without review. The rest, including all papers that we end up publishing, are sent out for peer review, to at least two and often more reviewers. We have about 80 staff serving the whole of Nature's editorial content, based internationally in half a dozen editorial offices which we maintain, of whom 25 are editors who are responsible for the assessment, peer review, selection and scientific improvement of primary research papers. (We have other editors responsible for sub-editing and production. Our selection processes also require administrative and electronic systems support.)

  These selecting editors are typically post-doctoral level scientists from excellent laboratories with strong publication track records. We select them not only for their academic abilities but also for their abilities to focus quickly and critically on key elements of scientific papers across a broad disciplinary range, and to engage with leading researchers in their fields. All editors are expected to travel internationally several times a year to visit key laboratories and meetings in order to keep abreast of their field and of the people in it.

  Nature papers are selected for their significant impact on a field. Here I use "impact" in its broadest scientific sense, rather that the narrow sense of indexes of scientific citations. The latter are high nevertheless as a consequence of our selection criteria. We seek to publish papers that significantly assist, enhance or, best of all, transform scientific understanding. The scientific impact may be deep (if a paper undermines or establishes a key scientific hypothesis for example) and/or broad (if it touches on several disciplines).

  We may exceptionally select papers because of their policy relevance even though they may not represent major new insights into the way the world works. Examples of this would be studies of uncertainties in climate models or predictive epidemiology of current disease outbreaks.

  It is sometimes falsely asserted that we publish papers to make headlines. It has always been my firm policy that we do no such thing. The frequent headlines that we already attract for our authors arise ultimately from the public's great interest in outstanding science. There would be nothing to gain by steering our selection towards the media in this way, and everything to lose in the respect of researchers.

1.   How do you ensure the integrity of the peer review process? Do you review your reviewers and, if so, by what process?

  We pay great attention to the quality of the peer reviewers that we use. Editors routinely track reviewers' publication output, assess the quality of reports that they receive, and explore the reputation of reviewers in conversations with their peers. We use many thousands of peer-reviewers from a very broad international base. I or senior colleagues will occasionally review the collection of reviewers used over a period within a particular field. We ask referees to disqualify themselves from refereeing if they have conflicts of interest, whether financial or competitive. We will also consider recommendations from authors as to who might or should not review their papers, though we are not bound by them.

2.   What are your views on the impact that a "pay-to-publish" business model would have on your peer review process?

  In principle, our process of peer review is independent of the sources of revenue. However, there are several considerations that, taken together, make me deeply uneasy about moving to an author-pays model.

  The first relates to editorial motivation, in my experience editors are highly motivated by the challenge of pleasing or stimulating readers—and what better measure of achievement than that individuals and institutions are willing to pay for the results? My colleagues spend their working lives adding value to the research process—in finding peer reviewers who will help authors improve a paper, in offering their own constructive suggestions, in detailed editing, and in selecting for impact. I do not welcome a system that removes the tangible reader-satisfaction feedback loop on our performance and editorial motivation.

  The second relates to economic power. I fully appreciate that a high-quality author-pays model has an incentive to maintain quality in the eyes of readers, but in the long run, being paid by authors rather than readers shifts the balance of economic power in a way that I would find extremely uncomfortable. In no areas of publishing other than listings, vanity and contract publishing do authors pay for publication. I do not maintain that current open access publishers are operating vanity publishing, because they too know that the reader is the ultimate judge. But shifting the economic power to the author cannot be assumed to be in the readers' best interests when it comes to the selection and editing of content.

  More objectively, our editorial process is expensive. I will leave it for publishers to debate the consequences of moving to author-pays, but there is absolutely no possibility in my view that an online-only author-pays model could hope to maintain the quality we achieve at numbers as small as $1,500 per paper.

  Another consideration relates to the long-term impacts on the selection process. Any publisher, whether commercial or otherwise, has to seek to maximize and develop revenue in order to develop as a service provider. If a publication cannot maximize revenue from subscribers, and with online advertising not being a strong revenue source, its publisher has little option but to accept more papers from paying authors. Thus there will unquestionably be a downward pressure on thresholds of acceptance, especially in those journals lacking diverse sources of revenue. There will also be a pressure implicitly to favour countries and scientists who can afford to pay more. Of course there would be no explicit discrimination, but one would have to be naive or disingenuous to maintain that there would be no hidden or unconscious pressures in these directions.

  And we editors value our financial viability. At Nature, I would far rather place my financial security on our ability to find hundreds of thousands of readers who (directly or through their institutions) will pay because we satisfy them, rather than a few thousand authors who will pay whatever their funding agencies will allow.

3.   Does Nature take responsibility for the papers that it publishes after publication? For example, in deciding to publish a paper, what weight is attached to the possibility that certain campaigners are likely to attach an exaggerated significance to certain papers which support their view? What steps do you take to respond to the distortion or misrepresentation of papers published in your journal in cases of great public interest?

  I have explained our fundamental criteria for selection above. It would in my view be misguided, both in principle and in practice, to treat anticipated misinterpretation or misappropriation of results as an additional criterion of our selection of papers. But we do take responsibility for the language used in papers and in our press releases, and we do regularly publish accompanying articles that explain the significance of the results and their limitations. In my experience, many authors and referees are duly sensitive to these concerns in assessing the presentation and interpretation of the results in our papers. It is a key role of editors to be on the alert for misrepresentations by authors, or for statements that could be misinterpreted, and to advise on rewriting if appropriate.

  We know that there is little more we could do to prevent the inevitable misinterpretation or misguided generalisations arising from papers that we publish. In such circumstances we tend not to get involved in the debate because (a) the authors or their peers will often be better placed to do so, and (b) it is important that we are, and are seen to be, impartial in our consideration of the scientific value of papers. For example, it is almost impossible to make any public statement about GM crops in the context of a particular debate without being instantly perceived as sympathetic to one side or another. It is of paramount importance that Nature be seen to be publishing papers on that topic for their intrinsic scientific or policy significance, not in support of other parties' agendas.

  All of that being said, whenever I see that our papers or procedures have been misrepresented, I will always consider responding to the media concerned. I or colleagues have responded on occasions.

4.   How does Nature implement the COPE guidelines on good publishing practice?

  Nature is not a member of COPE and therefore was not party to drawing up the guidelines. However, we do actively participate in, and sometimes stimulate, broader discussions about editorial practice, and I have discussed policies with my counterparts at journals such as Science and the Proceedings of the National Academy of Sciences. (It is because of such contacts and because we are a journal of basic research rather than medical research that we do not participate in COPE, whose role we nevertheless see as entirely positive.)

  Having read the COPE guidelines, I can endorse all of them and also state that we have been implementing them for years. The code requires that editors:

  Strive to meet the needs of readers and authors. Our high circulation and online access rates suggest we succeed in the former. Our Guide to Authors and recent innovations (see editorial in 18 March issue) exemplify our continuing commitment to author services;

  Constantly improve the journal. I hope that anyone comparing issues of Nature over the period 1996-2004 will see that this has been repeatedly happening;

  Ensure the accuracy of material we publish. As described elsewhere, we employ many staff to ensure this through peer-review and in-depth sub-editing;

  Maintain the integrity of the scientific record. Nature has always had a policy of implementing formal Corrections and Retractions promptly (see also Guide to Authors and appendix below). This often involves extensive negotiation with multiple authors, but we accept that burden. We also maintain ethical integrity by insisting on adherence to institutional and regulatory standards and having authors declare any competing financial interests. We also see it as our duty to pursue any accusations of author misconduct, and to bring established cases to the attention of readers;

  Ensure that business needs do not compromise intellectual standards. We find that Nature's commercial clients—advertisers and sponsors—value their association with us precisely because of our prestige, independence and integrity. Editorial independence is a foundation stone of all Nature journals; and

  Always be willing to publish corrections, clarifications, retractions and apologies when needed. As demonstrated in comments above and below, we are indeed willing.

5.   Can you outline your policy on retraction? What is a partial retraction, and how is this covered by the COPE guidelines?

  In the Annex are the relevant extract from our online Guide to Authors (see£9) and also the Nature and the Nature research journals corrections policy (see

  I have learned never to be too categorical about retractions because, to paraphrase Tolstoy, every retraction is unhappy in its own way, and one can lose sight of the needs of the community in debating definitions or prescribing protocols. Furthermore they are rare in our experience (about once per year seems typical).

  Bearing that preamble in mind, Nature's policy is that a formal retraction by definition applies to a whole paper, because its purpose is to remove the paper from the scientific record. Anything less than a full retraction is a Correction. If, say, authors had included in the original paper a speculative discussion that was not directly based on the evidence presented in that paper and that they subsequently wished to withdraw, they could do so by means of a formal Correction, which would then be linked to the original paper both electronically on our website and also in indexes such as PubMed.

March 2004

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2004
Prepared 20 July 2004