PLoS Medicine recently published Why Current Publishing Practices May Distort Science by Neal S Young, Omar Al-Ubaydli, and John Ioannidis. The article is facinating and I recommend it to anyone who publishes work in scientific journals.

Their argument starts with the observation that for researchers the ultimate accolade is to have work published in leading journals like Science and Nature because such journals turn down a high percentage of the manuscripts submitted to them. It follows that the leading jounals can promote themselves on the basis of their selectivity.

Therefore the leading journals are assumed to publish only the best scientific work. But Ioannidis et al argue that the reputations of these journals are pumped up by the same kind of artificial scarcity which is used to keep diamonds expensive. They go on to argue that this kind of scarcity can make it more likely that the leading journals will publish dramatic research which may eventually prove to be flawed.

With so many manuscripts chasing comparatively scarce pages in the leading journals, the manuscripts that are accepted could be the ones most likely to oversell themselves by presenting dramatic or important results which later are proven flawed. Less dramatic but possibly more accurate results are meantime relegated to less prominent journals.

Ioannidis et al base their argument on the fallibility of research published in leading journals on an earlier study of articles published in leading journals which were initially well-regarded, as reflected in citations by other researchers. However within 3 years about a third of the articles had been refuted by later studies. It should be noted that it hasn't yet been established that this isn't also the case for research published in less prominent journals, although their argument is supported by the economic concept of the winner's curse, which is the concept that someone who places the winning bid in an auction may have paid too much, because most bids are likely to cluster near the true value of the object for sale (envisage the bids forming a bell-curve.)

It's important to note that they don't suggest fraud, but rather that the leading journals have a potential flaw in the way they operate which makes it possible that flawed research will be published in them. It's interesting to consider to what degree the editors and peer reviewers of submissions can mitigate this problem, and its important to note that even if they can't do so the leading journals remain an important source of scientific information.

How does all this apply to Libertas Academica's open access journals? This problem shouldn't arise in OA journals because the marginal cost of publishing content is lower and circulation does not directly drive revenue. The basic principle we apply is that a paper is published if the peer reviewers and Editor in Chief recommend it based on the merits of the paper. The peer review critieria are just as robust as elsewhere but we don't face the same pressures as the leading journals do.