Ethical publishing (peer review journal articles in computational biology)

Following from my previous post about the PLoS ONE editorial board, I have been thinking about what might be involved in “ethical” publishing of research articles – and in particular about the choices that we as authors, reviewers or editors can make about which journals to support. Three main criteria spring to mind that could be used to determine ethical considerations:

1. What are the motivations of the publisher? In particular, are they a for-profit or not-for-profit organisation. Examples of the former include Macmillan, Elsevier and Springer; examples of the latter include not-for-profit publishers e.g. PLoS, university presses e.g. OUP, or learned societies e.g. AAAS or the Royal Society. What difference does this make? I can think of three:

(i) Does the profit motive have an influence on the articles published? I think in most cases (e.g. a bog-standard Springer or Elsevier journal, such as Journal of Molecular Evolution or Journal of Theoretical Biology), probably not. But in one notable case, Nature, the question does need to be asked as to whether the publishers seek articles with an appeal to the general news media, or with appeal to their popular science magazine (Scientific American).

(ii) Is it right for a for-profit publisher to benefit from pro-bono work of academic editors and reviewers? Our salaries are generally paid by not-for-profit organisations, such as universities and research institutes, and this funding in turn general comes from sources such as government, research funders (including charities) and student fees. Moreover, most of the publishers’ income comes from the same types of institution, whether through subscription charges or open access publication charges. Now, I appreciate that there are lots of benefits from editorial and revision work, including scholarship and reputation. But if we do such work for a for-profit organisation, are we being fair to our employers?

There is, however, a strong counter-argument. By using academic editors and reviewers who are not employed by the journal, the journal, and especially one run by a for-profit publishing house, can help ensure that its choice of articles to publish are motivated by scientific merit rather than commercial advantage. Thus there is a strong argument for editorial independence. Using the same examples as (i), JME and JTB have academic editors independent of the journal, but Nature have an internal editorial process to decide which articles to select for review.

A possible resolution might be for publishers to recognise the value of work put in on their behalf by employees of other institutions – institutions who are themselves customers of these publishers. For example, publishing houses could offer discounts off subscription or open access charges to employers of editors and referees on the basis of numbers of papers edited / refereed. This could actually add incentive to scientists to referee articles, which would benefit the publication process.

(iii) The profit motive means that money from publications is being used to generate wealth for the owners of the publishing house. A not-for-profit publisher has that money available to reinvest in the performance or dissemination of science. Which would I rather support? In the case of PLoS, this money is being reinvested into improving their publishing; a learned society might use this money in a variety of ways: if a journal is making a surplus, the money could support other activities such as conferences or fellowships.

2. Who has access to the published articles? Most scientific research (though not all) is funded by public money, or by charities. Should the results be available for all? Under the traditional model, institutions (universities, libraries etc.) pay subscriptions to publishers and receive journals (either on paper or electronically). This means that articles are only really available to people associated with such institutions. Excluded groups include private individuals, many health-care professionals, SMEs, and higher education or research institutions with insufficient funds, especially those in developing countries. Under the open access model, authors (or their employers) pay publication charges, and the published content is free for all. Increasingly, the societal (and admittedly economic) benefits of the open access model has been recognised organisations such as the UK government, the NIH and the Wellcome Trust. The argument is clear: results of research funded by public money should be available for the whole of the public.

This argument is compelling, and I am certainly glad that an increasing proportion of published research is available for all to read. There is, however, one obvious argument against open access: publication should be on scientific merit, and not decided by ability to pay. There are a number of ways in which this issue is being resolved. These include: (i) universities or research institutions having central pots of money to enable their staff to publish open access; in the long term this might be paid for out of savings by holding fewer subscription charges; (ii) funders of research providing funds for open access publication (e.g. Wellcome Trust); (iii) publishers themselves providing waivers to authors who cannot pay. For example, PLoS offer three scales of fees depending on the country of the authors, and publication from developing countries is free. This brings me to my third point:

3. Who has access to publishing in the journal? While open access ensures that research articles are available for all to read, the separate (and related) question is who has access to publishing work in a journal, assuming that the work is scientifically sound and clearly presented. Here, a number of issues arise:

(i) The most serious question is whether there are ‘privileged’ groups of authors who might be able to publish in a journal more easily than others. The most notable example is PNAS’s Contributed Submission scheme: PNAS state: “An Academy member may submit up to four of his or her own manuscripts for publication per year.” These articles are not subject to the heavy editorial filtering that takes place under Direct Submission, in which the majority of submitted manuscripts are not sent for review. Thus members of the NAS have a direct route to publish papers in this journal not available to non-members. Now, while NAS members are generally already at the top of their fields, and unlikely to gain much advantage from this, the people who really do get an unfair advantage are the students and post-docs of NAS members, who have access to publishing in a top journal, while the majority of students and post-docs do not have this advantage.

PNAS is at least transparent about its Direct Submission scheme. It is not clear how the editors of other top journals decide which manuscripts to send to review, and which are rejected at the editorial stage. Some journals, eg. Nature and Science, reject the vast majority of submissions at the editorial stage. For journals with editorial filtering, do the names and standings of the authors have an impact on the decision as to whether the articles are sent to review? Ideally, such editorial decisions should (and could) be made ‘blind’, i.e. the deciding editor is not given the names of the authors, and makes the decision purely on the basis of scientific content. I would very much welcome any such moves from journals with a strict editorial process.

(ii) The second question relates specifically to open access, as described above: might authors be excluded from publication purely on the basis of ability to pay. For authors employed by wealthy institutions, e.g. the UK’s Russell group of universities, this is unlikely to be an issue. But what about authors in smaller or poorer institutions, e.g. universities in the developing world, or authors in SMEs? Another pertinent point are authors who are still at school or who are undergraduates: this is unlikely to be the case in experimental biology, but does happen in computational / mathematical biology. Are such people able to publish in open access journals? Different journals offer different criteria: PLoS have a very clear policy with many waivers; OUP Journals offer free publication for developing countries, and limited waivers. BMC are less clear in their policies, offering waivers on a case-by-case basis. Of course, this issue does not arise with publication in journals operating under the traditional model.

In view of these thoughts, I have produced a table summarizing 25 journals that are relevant to my group’s research in computational biology – and publish articles in this area. These are by no means a complete list, but represent journals in which I have published (or submitted) papers, or from which I have read a reasonable number of articles. The journals are sorted first by whether or not they are not-for-profit, then on their level of open access, and finally on author accessibility. They are not ordered by my view of ethical concern. Rather, I have colour-coded each column separately using a green / amber / red scheme.

Journal Publisher Not for profit Open access Author accessibilty
PLoS ONE PLoS Yes Full Waiver scheme
PLoS Computational Biology PLoS Yes Full Waiver scheme. Non-blind editorial filtering
Nucleic Acids Research OUP Journals Yes Full Limited Waivers / Developing countries free
Bioinformatics OUP Journals Yes Optional Yes
Briefings in Bioinformatics OUP Journals Yes Optional Yes
Journal of the Royal Society Interface Royal Society Yes Optional Non-blind editorial filtering
Genome Research CSH Press Yes After 6 months or optional immediately Non-blind editorial filtering
PNAS NAS Yes After 6 months Contributed submissions. Non-blind editorial filtering
Artifical Life MIT Press Yes No Yes
Molecular Biosystems Royal Society of Chemistry Yes No Yes
Science AAAS Yes No Non-blind editorial filtering
Proc Roy Soc B Royal Society Yes No Non-blind editorial filtering
BMC Bioinformatics Springer No Full Limited waivers
Genome Biology Springer No Full Limited waivers
BMC Systems Biology Springer No Full Limited waivers
BMC Genomics Springer No Full Limited waivers
Molecular Systems Biology Macmillan No Full Exceptional circumstances only. Non-blind editorial filtering
Journal of Theoretical Biology Elsevier No Optional Yes
Mathematical Biosciences Elsevier No Optional Yes
Bulletin of Mathematical Biology Springer No Optional Yes
Journal of Molecular Evolution Springer No Optional Yes
Journal of Mathematical Biology Springer No No Yes
Journal of Computational Biology Mary Ann Liebert No No Yes
Biophysical Journal Elsevier No No Yes
Nature Macmillan No No Internal non-blind editorial filtering