Thursday, June 27, 2013

Open Access - an insider's view

This is my mum's (Alice Meadows') response to my post on Open Access. She works for Wiley.

Your post hits on many of the key issues in open access (OA) and also raises some interesting points.

You are quite right that scientists may choose to publish OA so that more people can access their research. However, although it does seem to be true - unsurprisingly! - that OA articles are downloaded more and reach a broader audience, it is not yet clear that this leads to more citations. There have been several studies of this including one by my fellow Scholarly Kitchen blogger, Phil Davis, which indicate that OA articles are not, in fact, cited more. This could well change in future, of course, especially as more - and better - OA journals are launched. But for now the jury is out…

You are also right to point out that paying to publish will not necessarily lead to reduced costs for institutions (and I say institutions rather than libraries advisedly - it is by no means certain that this budget would stay within the library).  Your back-of-the-envelope calculation about the cost to UMass of moving to a fully 'gold' model (where you pay an article publication charge or APC, like PLoS) is echoed by, for example, the University of Oxford which has calculated that the APC costs for all its research output would likely exceed its current subscription costs (and would certainly exceed the amount it receives in its block grant from the Research Councils UK for APCs).  Other, less research intensive institutions, would certainly benefit financially from a move to Gold OA though - as would commercial organizations, such as pharmaceutical companies, who would have free access to large amounts of research that they currently have to pay for. There is also an alternative to gold OA; green OA permits an author to make a version of her/his article freely available in a repository, such as PubMedCentral, typically after an embargo period of somewhere between 6-24 months.

Both versions of OA have their pros and cons.  Gold OA acknowledges that there is a cost to publishing, which is a good thing - not just from my perspective as a publisher, but from the perspective of everyone involved in the scholarly communication chain, all of whom are adding value in one way or another (that's why we get paid!).  There really is no such thing as a free lunch, so trying to make research articles completely free - no matter how appealing that may be philosophically - is not a realistic option. Someone, somewhere will be paying for it, either overtly (eg through advertising or APCs) or covertly (eg, by selling personal data to other companies or by spending time publishing instead of researching).  However, there is huge disagreement about how much it costs to publish an article, from the $99 PeerJ membership model through to the $5200 it costs to publish in Nature.  And, as they say, if you pay peanuts you get monkeys (interestingly, PeerJ's logo is a monkey!) so your point about the quality of editing (including peer review) in OA journals is well made.  The last thing anyone wants is to have bad science freely available.  Peer review is not a perfect system, but most scientists believe it works most of the time. I don't believe that PLoS One is publishing bad science - as I understand it, their 'light' peer review simply means that reviewers don't assess articles based on novelty or fit, but only on whether they are scientifically sound. That enables them to charge considerably less for PLoS One ($1350)than for, say PLoS Biology ($2900), which has a more intensive peer review process, taking into account scientific importance as well as rigor. You also noted that not everyone can afford to pay an APC anyway - some disciplines, such as mathematics and physics, receive relatively little funding, and there are large parts of the world that can't afford to pay to publish at all (though many publishers reduce their prices in those cases). But lack of money should not disqualify researchers from publishing - and from publishing in the journal of their choice, not just the journal they can afford.  Nor should the relative cost of an APC be used to discourage scientists from publishing in their preferred journal.

Green OA is sometimes described as 'no one pays', however, it is becoming increasingly apparent that someone does (need to) pay, because – for now at least – green OA relies on the continuation of the subscription model to survive.  This is why the issue of embargo dates is so important - too short an embargo and there's a risk that libraries will cancel their subscription; too long and the whole point of making an article publicly available is lost. The jury is out on this issue too - no one really knows how short is too short, though it does seem very likely that this also varies by discipline. And how do you figure that out without jeopardizing the existence of the journals involved, not to mention the scholarly and scientific societies that support them?  One possible method is to use half life analysis, looking at how long it takes before an article has achieved half of all its lifetime views, which could help ascertain a 'safe' embargo period that will protect journal subscriptions. Another option is to allow authors to post an early version of their manuscript from the get go but, as you point out, that carries the risk of bad science escaping onto the web. Although scientists may understand the distinction between the accepted version of a paper and the final version, the general public probably does not.

My own personal feeling is that there is a lot to love about OA. It's revolutionizing all of scholarly communications and forcing everyone who is involved to think again about things we've long taken for granted - Impact Factors, tenure, peer review, publishing ethics, and more.  But at the same time, I worry that a little bit of knowledge can be a dangerous thing. The OA debate so far has been dominated by a relatively small number of loud voices from a relatively small number of disciplines and organizations, who are influencing some awfully big decisions that could have a significant impact on science and scientists, such as what license an author must sign (do all authors really want to sign away all their rights under a Creative Commons By Attribution (CC-BY) license?), how long an embargo period should be, whether green or gold OA is better.  We don't have all the answers to these questions yet and we need to do more to ensure that everyone's voice is heard in what I hope, going forward, will be a less polarized and more collaborative approach to solving these issues.

These views are my own and are not necessarily those of my employer


  1. I think the observation that open access articles get more downloads, but this doesn't always translate into more citations for two reasons. The first is that more downloads doesn't equal more readers. Especially when I am off-campus, I for one tend to download open access articles multiple times as I read and refer to them (because I can), but don't do the same for subscription-only ones (some of which imposed convoluted access rules). So for the subscription articles, I will download or print the article and refer to the one copy. The second reason why more downloads doesn't equate to more citations could be because more members of the public are reading it (or more people who are not necessarily writing scholarly papers). Which is a reason why we need more complex metrics of "success" for our publications.

  2. Interesting! The first reason is what I think of as a 'bad' reason as it doesn't reflect a genuine increase in usage, whereas the second would be a 'good' reason, with more usage from a wider readership.