Earlier today, the British Academy released a research project report, Open access journals in Humanities and Social Science. The project was funded by the Higher Education Funding Council for England (HEFCE), and was meant to address practical issues that may arise surrounding open access (OA) publishing, particularly in the Humanities and Social Sciences (HSS). There is a significant lack of data used to back up a substantial number of arguments and suggestions for policy surrounding OA. Part of the goal of the project was to collect data that can be used to help drive evidence-based decisions.
Some key conclusions of the study:
- Gold OA is unlikely to play much of a role in HSS disciplines, Green OA is much more feasible
- There is a good amount of variance in the types of publications used by different HSS fields (book chapters, books and journal articles). There is a significant amount of publication that takes place outside of the UK, and publisher compliance with UK OA regulations varies quite a bit (though larger, multi-national publishers generally do better with compliance). Because of this, “The current rules for RCUK grant-funded publication…at present make non-UK journal publishing in the Humanities very difficult, and in literature/art/music-based disciplines almost impossible”
- The study looked at article usage half-lives, and unlike previous studies, included older archived content from JSTOR. The data shows that article half-lives are likely longer than previously suggested. A 1:2 ratio for embargo period lengths is concluded to be appropriate, but the dividing point should not be STEM:HSS, rather given the actual usage patterns of articles, it should be Medicine (1): HSS, Physics, Mathematics, Chemistry and Life Sciences (2). Suggested embargo lengths are 12 months (Biomedicine) and 24 months (all other fields).
- Survey data from librarians states that embargo length does not play a role in decisions to retain or drop journal subscriptions. This is because readers want access to the Version of Record (VoR) for papers, rather than the Author’s Accepted Manuscript (AAM), and because Green OA compliance is so haphazard. Papers are scattered, hard to find, and because of low compliance, often unavailable, which makes subscriptions necessary.
- The core issues about the costs of publication must be separated from questions about access. A key to improving the cost situation is to remove the current information asymmetry, and require more transparency on the part of publishers to match the public availability of library budget information.
In the interests of transparency, I’d first like to clear up any questions of conflict of interest, both in this interview and in the study in general. You do have an ongoing relationship with Oxford University Press (OUP, my employer). Did this in any way impact the study?
I am indeed a Delegate of OUP (i.e. on its management committee); I have several other potential conflicts of interest too, such as being involved in several journals and in the UK research evaluation (REF) exercise. This didn’t, however, affect what was principally driving me: a growing sense of fury at the fact that major decisions were being taken about OA in Britain without serious consideration of their effects, because people didn’t know or understand what these might be.
When we – myself and my two researchers and co-writers, Rebecca Darley and Dan Reynolds – were doing the research, other publishers sometimes expressed worry that I might not be neutral because I was linked to one of their rivals, and people in other sectors were equally concerned that I was simply the publishers’ spokesperson. As to the first: I consulted OUP as much as and no more than I did five other major publishers. As to the second: anyone who reads the report will not, I believe, easily think that we are arguing the publishers’ side. Not that the report is particularly polemical; it’s aimed at facts. But the data we set out do not favour any particular side, and in fact undermine arguments made by each of the many sides.
The report talks about the intense “heat” of the OA argument, but about the scarcity of data offered to back up arguments from most directions. In particular you cite the RCUK (the UK Research Councils) and other public bodies for, “making assumptions about the likely future open access policies of non-UK journals which were based on very little evidence.” What can governments, publishers, advocates, funders and librarians do to help improve the decision-making process so it is more evidence-based?
I found in early discussions with RCUK that they were running their assumptions on very optimistic readings of the policies of quite a small number of major journals, and assuming that these stood for the whole. This is not the case, as we found. So one answer to your question is, I guess, simply do more studies like this one. And not just for HSS: for STEM, Natural Sciences and Medicine, too, which public bodies tend to assume are unproblematic because they are moving to OA faster, but that is not fully the case.
Actually, there are several useful studies of individual aspects of OA. Phil Davis’s work, which came out just before Christmas and which is flagged in our report too, is an important starting point for usage half-lives, for example; the EU’s PEER study is good too. Conversely, the JSTOR part of our own study was important for our understanding of how researchers actually use journal articles over the long term, but incomplete; a fuller study of content providers of the JSTOR type could and should be extended to Natural Sciences and Medicine too, as the Davis study does for publishers’ half-life figures. Players in this debate don’t understand well enough how any part of the sector manages its journal publishing and journal reading. I hope our study will contribute to a better understanding of how it works in HSS, at least, but there is quite a long way to go.
Similarly, you note the lack of granularity in policy, a lack of recognition that there are differences between different disciplines that must be addressed. The focus of this study is the Humanities and Social Sciences research communities. Have the needs of HSS researchers been ignored in both policy and the OA debate in general?
In every country, policy-makers start with STEM first and assume that this stands for the whole. In a way it is natural; Medicine, Physics etc. are so expensive that they are the most visible part of the sector, and the public rhetoric of investment in science has these disciplines in mind too. I don’t mind that. What I mind is the forcing of half the researchers in each country (that is the rough percentage of HSS pretty much everywhere) into boxes which are not appropriate for them. It is not actually that hard to think in a more granular manner from the outset, to avoid the inevitable push-back later from our half of the sector and the resultant compromises; if you don’t do it from the start, it takes more work for both sides.
You state that Gold OA is unlikely to play much of a role for HSS researchers. What are the key differences and what approaches are available that best meet HSS needs?
STEM disciplines are in many cases more used to Gold; grant budgets can pay for article processing charges (APCs), and in Medicine often have to; in many disciplines, people submitting articles had already to pay up front in various ways as well. That has not been the case in most of HSS (Psychology is a major exception, half in Science as it is); and, furthermore, a far smaller part of research work is based on external grants. One well-known aspect of the UK grants culture is that HSS, covering 50% of researchers, gets 10% of national public research money. Nor is that particularly unfair; archival or text-based work by a single researcher costs little, and even large-scale collaborative research with us seldom needs much plant. But of course it also means that the great bulk of research is not grant-based; so there is no one to pay for APCs; so only a few contributors to any HSS journal – I would reckon under 10%, even in the medium term, although I haven’t done the research here to show it – are ever going to ask for Gold OA. Unless they are forced to, which is not the case in the UK or (still less) anywhere else. So, if you want to push OA in HSS, it’s got to be Green. Hence the fact that embargo periods are a particularly hot issue in HSS.
The study also shows the significant amount of research publishing that is done by, and owned by the research community itself through societies. Has a focus on the profit margins of larger commercial publishers obscured the role of the research society, and do current policies threaten their existence?
Research societies vary a lot. They are non-profit, and they put their surplus from publishing back into their disciplines by funding early-career researchers, conferences etc; that makes them similar in one crucial and valuable respect; but the way they do it is hugely variable. It is like the late Cretaceous, with dinosaurs and mammals: some of the big Science or Pharma societies behave like the most notorious publishers, charging amazing amounts for journals; others (most of those in HSS) are tiny operations, run by people who sometimes barely understand what is hitting them right now.
I don’t think there is anyone who doubts that the role of all kinds of research society is of major value for the academic sector as a whole. I do think, however, that, again, policy-makers have the big STEM societies in their mind when they say that it shouldn’t be journal buyers who pay for the work research societies do. And I also think that this shows up something which underlays the thrust of our report as a whole: that in every discipline (but differently in each), the research ecosystem is all tied together through very subtle and complex relationships which are also financial; and if you try to affect one element (e.g. by insisting on certain rigid forms of OA) you risk affecting the others in unpredictable ways. Unpredictable, at least, if you don’t do the research.
My sense of the UK policy is that it was designed to put the UK in a leadership position, that the RCUK’s policies would set a standard for the rest of the world to follow. Given that most other countries have now shown a clear preference for Green OA over the RCUK’s favored Gold OA, is the RCUK’s policy still viable? If the UK remains on this course, how will this affect the international standing of UK researchers?
Yes, I am sure you are right about UK policies. Actually, Gold is getting more interest now in central Europe in particular – the Austrians, for example, are experimenting with Gold monographs – so the UK isn’t quite as isolated as it seemed a year ago. Medicine might eventually go Gold overall, too. And some new but respectable publishing houses are now offering very low APCs for journals, and getting more interest. But there is little sign that anything except Green is going to dominate world-wide in the medium term. The UK is thus in the first-adopter position, and that is always risky.
It then depends how UK public bodies act. HEFCE, the main funding body for UK University research (which also funded our report), has said in its OA manifesto for the next REF, which is just out, that it has no preference at all for Green or Gold, and is also happy with 24-month embargoes for HSS; it also has a generous let-out clause for anyone whose research is best (‘most appropriate’, the text says) published in a non-OA journal. That is a position which I am quite happy with, as long as it is not misinterpreted by University research managers – which is, sadly, far from inconceivable. RCUK is having a consultation this year and has the opportunity to do the same, stepping back from its somewhat restrictive present policies. I hope it will! But it’s important to stress that HEFCE is more important for the majority of HSS researchers, who do their work without grants.
Your study offers a fairly straightforward approach to research embargoes, putting Medical research at a 12 month embargo, and all other fields at 24 months. Can you explain the rationale behind this, and why you think it’s both fair and viable?
The rationale is that usage half-lives – the Davis report made this clearest – are far shorter for Medicine than for anything else. Mathematics and even Physics are up there with HSS, at around twice that of Medicine. The clamour for short embargoes is also above all coming from Biomedicine. But embargoes, it must be remembered, are an attempt to deal with two opposing pressures: the desire by some researchers to have full access to all research as soon as possible, and the felt need to protect journals from going under because no-one needs to buy them because all the research is already free. Usage half-life is a different issue: it is the analysis of when it is that half of the downloads of each article, averaged across each journal and then each discipline, have been made. So one is a political/moral/financial battleground; the other is simply an observed set of data. They are not fully commensurable, and you can’t read one off into the other without considerable nuancing. Half-lives are about double these embargo figures, in fact, but that doesn’t necessarily mean that embargoes should be longer; that depends on many other factors.
The observed half-life contrast between Medicine and the rest is the reason why we argued for that division in terms of disciplines, then. We went for 12:24 in embargo terms simply because these embargo lengths are where the debate has been concentrated; and also because we couldn’t see why it would really be necessary for everything to be OA in Medicine already after 6 months (the current UK rule) when half-lives even in Medicine were four times that – i.e. it really isn’t the case that all Biomedical researchers are reading everything in the first week it comes out and that it goes out of date in week 2, which is the impression that some of the rhetoric gives. That is not a fully adequate reply – and I know that you, David, have argued that embargo periods need to be looked at a lot more carefully (and granularly) than they have been in order to get them right. But Medicine vs the rest seems to me a firm starting-point.
The question of embargoes, and their potential impact on journal subscriptions has been the subject of much debate on this blog (here, here and here for example). In general, the librarian response has been similar to what your study saw, that embargoes do not play much of a role in subscription decisions because Green OA is so disorganized, and availability of papers is somewhat random, with poor discoverability and low compliance rates from authors depositing in repositories. But what happens when Green OA policies, like that of the US Government, become the norm? Systems like CHORUS are being built with the goal of making Green OA automatic, greatly increasing compliance and discoverability. If Green OA becomes comprehensive and systematic, will this change its impact on journal subscriptions?
The US, and EU initiatives favouring Green are only for grant-funded research. It’s only in the UK where the initiative covers all of research, because all research is measured in public research exercises which have a real effect on University funding. Outside the UK, HSS research, in particular (and also Mathematics) will mostly not be covered by mandated OA policies. Under these circumstances, the bulk of HSS journal articles will only be posted in a Green context if their authors specifically want to. Experience shows that this posting is at best intermittent (even in places like Harvard which try to enforce it). So libraries will still have a hard time cancelling journals, for reasons set out in the articles you cite, and also in our own report. If Green posting became as normal in HSS as in Physics (which I have a hard time believing will be the case), that would certainly change matters. All the same, it is worth stressing that the EU’s PEER survey also showed that even Medics were insistent that they needed to see final published versions as well as posts, because so much might change between peer-reviewed text and final publication; and HSS academics, with longer articles, tend to want to cite page-references and the like as well. Only Physicists were more relaxed about Green posting in that study. I know there is a fierce argument about whether arXiv really has had an effect on Physics journals, but it certainly hasn’t had a total effect, and arXiv has been going for a good while; in HSS I don’t see it any time soon at all.
Your study cites information asymmetry at the core of the problematic relationship between libraries and publishers regarding the cost of publication. If, as some have suggested, this is an irrational market, how would transparency lead to change?
Well, I’m always optimistic. Certainly, it is hard to bargain with a publisher if you do not know how much it can afford to accept as a subscription, but the publisher knows what your budget is because it is publicly available. Like buying a carpet in Turkey or Morocco: you can’t bargain if the seller knows what your real top price is and you don’t know his real lowest one. Transparency levels the playing-field.
The article you cite does not however really focus on what I see as the crucial issue – how can University library buyers (overwhelmingly the main buyers of academic journals) – actually have a proper bargaining relationship with publishers? Publishers often try to pick them off one by one; also – I am told – US cartel laws make it harder for libraries to get together there to negotiate terms. But it’s most logical for collective bargaining to take place here, on the basis of real information. And, in terms of how library finances are, this seems to me more important than any OA policy at all, which is more like water dripping on a stone. Effective, possibly, but certainly slow.
How would you like to see the UK government use this study? What changes in policy do you think should result?
I’d like RCUK to be more relaxed about what they require in OA terms from grant-holders; what they are doing now may not matter in some disciplines, but it certainly matters in much of HSS, and it will have very negative effects in literary disciplines.
Apart from that, I’d like government(s) to have more granular academic policies from the outset, rather than us constantly having to fight to get the position of HSS disciplines recognised for what it is, in each case. But that isn’t the theme of our report.
Was there anything that came up in the course of your research that you found particularly surprising?
Yes, I had no idea that usage half-lives would turn out to be so uniform, even across HSS, never mind STEM as well (apart from Medicine). I thought the figures were wrong; I tried to use different methods to assess them; I tried JSTOR to see if the pattern was different (which it is in many ways, but not this one). But I concluded that the data were firm, and that that was how it is.
How do you see the future of scholarly publishing? Are we likely to settle on common standards for access policies, or will the future see a mixed marketplace of different approaches for different fields from different governments and funders? Will Gold, Green or Subscription Access models dominate?
God knows. I have no doubt that OA will get steadily more important. Will it ever replace subscription models for HSS? I am inclined to doubt it. So, overall, I see a mixed marketplace emerging, with no common standards. And other countries are currently all going in different directions: state-funded OA in Latin America, content-providers in the USA and France, state-enforced journal hierarchies (which privilege traditional journals and their market advantage) in Italy, and so on. The usual muddy picture, in fact. Not a place for starry-eyed visionaries – but perhaps harsher for inflexible traditionalists.