Summary

Advantage

Comment

Everyone with access to the internet would have immediate access to all published papers without having to pay.

This is highly desirable, but dissemination of the information will have to be paid for by someone. An economic comparison of the different models is needed.

Note: copyright is not an issue; journals need not require authors to cede copyright, even under the current system; a licence to publish should be sufficient.

Open access publishing will boost citations.

An argumentum ad invidiam, targeting everyone’s prejudices about citations and impact factors. And there is no evidence that this will happen. Even if it did, it is hard to see how science (as opposed to those taking part in the RAE race) will benefit by increased citation rates.

Disadvantage

Comment

Ex nihilo nihil fit; someone has to pay.

The average cost of publishing an open access paper has been estimated at $5000-10000. The author pays model will involve either a submission fee and a publication fee or just the latter, but at a higher price. Most authors do not want to pay to submit articles to journals or to have them published; those who are prepared to pay do not want to pay more than $500 a shot. How many shots will it take to publish a paper if you have to run the gamut of having it rejected by high impact factor journals, each of which charges a submission fee? Can you or your institution or your grant-giving body afford it, without disadvantaging the funding of other research? [The Proceedings of the National Academy of Sciences has experimentally proposed a voluntary surcharge for optional open access publishing; the charge has been reduced from $1000 to $750 in 2005; probably there wasn’t enough uptake at the higher charge; I predict that the charge will be reduced further, if the system is not abandoned instead.]

The system will advantage those who can afford it.

Large pharmaceutical companies, for example, will be more readily able to publish than academic institutions.

The system will be open to abuse.

The wide availability of the internet means that open-access journals will spring up everywhere and vanity publishing in science will be possible. Non-peer-reviewed papers will multiply. As throughout the web, distinguishing the wheat from the chaff will be difficult.

Peer review will suffer.

Currently most peer review is carried out without payment, relying on reviewers’ goodwill and sense of duty. The system is flawed, but as good as we are likely to get. If authors are expected to pay for publication, they will expect to be paid as reviewers, increasing costs.

Developing countries have limited funds and more limited access to the internet; they will be disadvantaged; so will impoverished researchers elsewhere.

Some grant-giving bodies will add 1-2% to their grants to fund author pays publishing, but that will not cover research that is not so funded. Research that has no formal funding to cover publishing costs includes most of the research that is carried out in developing countries. Important work in developing countries, tackling practical local issues that will not be researched elsewhere, will suffer. Computer-based or library-based research, such as reviewing and systematic reviewing, which needs little funding, and a lot of opportunistic bedside research will suffer. Students, both undergraduates and postgraduates, will have difficulty in publishing their work if they do not have support from a senior scientist with funds. Junior hospital doctors will also be disadvantaged. Subsidizing those who cannot pay will increase the costs to others. [The Public Library of Science, with its $9M grant from the Betty Moore Foundation, is offering free publication in PloS Biology without question to anyone who wants to take advantage of the offer; will you be able to resist?]

In some countries PhD theses are bound volumes of published papers.

Such theses will be prohibitively expensive to produce.

In an open market, journals with high impact factors will charge authors more for the privilege of publishing, taking advantage of the inevitable competition that will be created by research assessment exercises.

Alternatively, journals with low impact factors may seek to increase submissions in other ways.

"This week’s special offer: all papers on treacle receptors in tropical frostbite submitted to the Journal of Indeterminate Results will be charged the special submission fee of $100."

"Submit one, submit one free."

Some grant-giving bodies may insist on publication in open-access journals as a condition of funding.

This will increase the power that a few individuals can exert over researchers.

Not-for-profit publishers, such as learned societies that rely for their scholarly activities on income from publishing learned journals, will be disadvantaged.

Some may go to the wall. It is not encouraging that the BMJ Publishing Group is now charging for access to online material on bmj.com, because of dwindling revenue from subscriptions and advertising, following the introduction of open access (www.biomedcentral.com/openaccess/archive/?page=features&issue=13).

Learned societies that publish journals also currently benefit from sales of reprints, for example to pharmaceutical companies. Open access will erode this source of income.

There is a question about the quality of open access journals.

In my own field the publication record of the open access journal BMC Clinical Pharmacology is poor: it published four papers in volume 1 (2001), eight in volume 2 (2002), four in volume 3 (2003), and six in volume 4 (2004). In contrast, subscription journals in the field are growing. Universities under the lash of the RAE demand publication in journals with high impact factors: BMC Clinical Pharmacology doesn’t have one and the impact factors of other BMC journals are as yet disappointingly low; perhaps they will increase acceptably with time.