The big news in scientific publishing this week is that the NIH has gone open access. I’m sure that almost everyone on the scientific blogosphere will be saying how wonderful it is that research has to be made public, and the advocates for open access will be happy. So, I’ll take a slightly contrary position. Whilst open access is certainly good in principle, will it work so well in practice? I’m not 100% sure, I can see some problems.
To start with, let’s note that publishing scientific papers cost money. A few years ago I heard an estimate that a paper would cost at least $100 to be published. This was the bare minimum: just for the editorial process. The charges for publishing are much greater, some running into a couple of thousand dollars. One reason for the differences in price will be the cost of printing the journal on real paper, but other costs have to be paid, such as advertising the journal, and overheads (e.g. to pay for the web site), as well as some profit and sending the editors to exotic place for meetings. So where does this money come from?
There are two primary sources of money for journals: either the reader or the writer pays. Most journals have a reader pays policy, otherwise known as “screwing as much as you can out of the libraries”. We (scientists) have to have access to a lot of journals, so the publishers can take advantage of the scarcity and really push the price up. Looking at direct costs, the libraries lose but the scientists gain, because we don’t have to pay for our work to be published1.
The alternative is that the writer pays. This then allows the journal the opportunity to let the reader see the publication for free. Obviously this is much easier with electronic access: there are no printing costs. Hence, it allows the development of open access. Again, looking at direct costs, this is good for the libraries (well, almost – see below), but not for the scientists, who have to fork out the large sums of money themselves. Which means they need to have money available to do this. Sometimes they won’t.
I can see three situations when money may not be available to pay for publishing. Firstly, the scientists may be working in the developing world, where there is little funding available for science anyway. In this case, the authors can apply for a waiver, and I think it’s routine that the journals expect to grant a certain proportion of waivers for precisely these people. Of course, the other authors pay instead, but overall I think that this is a good thing.
Secondly, a scientist may be in a rich Western country but not have any research money. Sometimes their funding only covers their salary, or they may not be employed. In this case, a journal might have less sympathy, simply because the author is from a rich country. And if they are generous with their waivers, then authors might notice, and start asking for them when they would have access to funds (e.g. if they are part of a group that has general funding, but they are only receiving a student stipend). This can be ameliorated to some extent by shifting funding patterns so that either salaries come with money for publication etc., or journal costs are paid from the department or university, so that the costs are recovered from overheads.
The third case is, I think, the most interesting. Even successful groups have finite funds. If one project is particularly successful and produces a lot of papers, it has to pay out more money. It will have budgeted for a certain amount of publication cost, but finding the extra money might be difficult: it has all been spent on producing the results or publishing the early manuscripts. Shifting payment to the department only helps slightly: it just moves the problem up one level.
So, open access is not without its problems. But there might be solutions as well. One would be for universities (or, in practice, their libraries) to pay a fixed annual fee and then allow authors to submit without paying. This then shifts the fiscal responsibility back to the libraries.
What of the journals? Under the old model, they have scarcity on their side: researchers need access to a wide range of journals. Under the new model, it is less clear. A top journal like Nature still has scarcity: we all want to publish there, so we will be prepared to pay (I guess this is one form of vanity publishing). But for the mass of journals, they have a lot of competition. Don’t like The Journal of Applied Felinology? Then submit to the Journal of Applied Cat and Kitten Studies. For these journals, the scarcity shifts from them to the scientists: they have to compete to get our good papers. They did anyway, but now we also have the money, and hence the opportunity to laugh evilly as we make them compete for our latest dissertation on the release of calcium. We’re usually short of funding, so we will consider the price when deciding what journal to submit to. This should push the price of submission down, thanks to the glories of the free market. The publishers therefore don’t become fat at our expense, and we might get a better service.
Except that scientists are cheap-skates and have plenty of things to spend our money on. If we don’t have to pay for getting a paper published, why should we? And a lot of journals will happily publish our papers for free (and charge the reader, of course). So, open access journals are competing for papers at a disadvantage, because they are expensive. They’re relying on scientists having a lot of money (and not all of use work in medicine), and also having a sense of idealism about open access. This suggests to me that open access is unlikely to take over scientific publishing, and will do better in some areas than others – depending in part on how rich the area is. For example, compare BMC Ecology to BMC Evolutionary Biology. BMC ecology published three articles in March, whereas BMC Evolutionary Biology published 27. I suspect this can be partly explained by there being more money in evolutionary biology, especially as many of the papers are based on molecular work, which is expensive. I suspect this pattern generalises, but I’m too lazy to look any harder. Plus, I have to continue my felinology studies, into the effects of squeaking on the service of food. I think I’m the one doing the study.
1 OK, some journals charge the author as well. Often this is for long papers, or for colour pages. But some, mainly in the US, do charge for anything. One group of journals not only charges, but also manages to provide a bad service: over a year from acceptance to publication.
When one of my papers was published in an open access, online-only journal, my PI didn’t pay at all; our institute has a membership / subscription that covered all publication costs. This was a BioMed Central journal, I don’t know if others have similar arrangements.
I thought we had the same thing, but I checked and we only get 15% off. That might be the way things will go in the future, but I don’t know what the economics will be like. It’s a curious problem.
Really good post, Bob. As you can imagine we discuss this sort of thing all the time in the Nature
orificeclosetpalatial suite, but unlikepoliticsreligion“classic radio comedy”:http://en.wikipedia.org/wiki/Round_the_Hornepets, I don’t dare pronounce on such things, leaving such matters to my distinguished superiors. (Aren’t strikeouts great? Now I’d like to know how to write in Scarlet Letters.This is an excellent post Bob. You cover the problems nicely, although I’d like to see some easy solutions (wouldn’t we all?) I’d also like to increment your comment count for a serious post.
Bob, depending on your area of research, the publication cost will be a small percentage of the overall cost of the paper (if you consider both material and salaries). The bigger problem for me is that – with the exception of PLOS – the better, or at least more respected journals are usually not open access. So I try to submit my research work to the journal with the highest Impact. My next job or grant might depend on this.
@Henry – “distinguished superior” ?! [not! in case anyone was wondering] Oh well, beats being a damsel in distress, I suppose (marginally).
@Bob, you write a very good post with many thoughtful arguments. At risk of sounding like Mandy-Rice Davies (“he would say that, wouldn’t he”) and of bringing down the wrath of the
Spanish Inquisitionevangelists, I venture a few responses to the issues you raise from the Nature editorial perspective, in the spirit of a rational conversation.Costs. Costs and access are publisher issues, not editor. Probably editors would make everything free because they love to have their material read, then very soon they wouldn’t have a publication in which to publish.
Open access and subscription are two business models (See Nature’s web focuses here and previously, which went into all this in some detail from a range of perspectives — I have yet to read the details of a successful OA business model. Despite one or two recent claims I have seen no detailed case-study, and the OA publisher with the highest impact journal has a large charitable bequest. (This is not intended as a critical comment, but as a fact.)
The costs of producing a paper are far, far greater than the OA charges at the list to which you link, and are only intended, I believe, as a contribution. (See here for article by Martin Frank of the APS). The peer-review process itself is expensive, especially if the journal hires professional editors and runs a rigorous system. (_Nature_ for example has about 30 such editors and publishes around 20 papers a week.) Print is a cost, but a professional web publishing is also a costly exercise. Not only the maintenance of the articles themselves, but meta-services such as manuscript tracking systems that have to be maintained and developed.
Journals also add much value to papers, both in online added value and editorially. All this costs. (Authors in Nature love to have news and views articles on their papers, or editors’ summaries, a press release, the cover; as well as lots of extra features online).
Your point about authors paying for OA — most journals that operate the OA system have a blinding process so the editors are not aware of whether the authors have paid, had a discount or waived. To me, with my editor hat on, the question of payment for publication at all is a wedge that has a thin end to it. The subscription model does have the advantage of a demonstrable level playing field for all submissions.
Do you know about the UN programmes HINARI, AGORA and OARE? Many publishers, including NPG, are part of these programmes in which all their published scientific papers are made free online to participating countries, of which there are many. You can follow links at the bottom of nature.com web pages if you are interested to follow this up.
There are a lot of interesting issues raised in your post: I hope that some of these points I’ve mentioned are of reciprocal interest.
@Richard – stop taunting me, you brazen hussy.
Couple of comments. One is that in actual fact the majority of existing open access journals do not charge for access. However, as you would expect these are mostly journals with relatively low impact factors with specialised audiences, run by learned societies. The more contentious statement on that link is that 75% of non-OA journals charge author side access fees. This is partly because society journals have moved quicker to open access models because their business models make it easier I think, leaving most non-OA being commercial publishers (that’s my perception, not backed up by numbers, the lines get blurred in many cases anyway).
I think the point is that this overall ought to be a zero-sum game. Someone is already paying for the process. Researchers already pay for this both directly and indirectly. Publishing in PLoS Biology (~£1000 I think) can be cheaper than paying the page charges in some other journals (I paid slighlty more for colour and page charges for a paper a few years back in a high impact toll access journal than I did for a recent paper in PLoS ONE). We also pay through taxes (call it the university direct grant, call it full economic costs, whatever) which get top sliced and go straight to supporting libraries which are charging us more for being able to access less journals because they cost too much (is this different in the states, what is the funding model for US university libraries?). We are also paying increasing charges for the re-use of material (I paid recently to reproduce a figure in a review). If all of this gets wrapped up into a transparent publication charge I think things will be much clearer.
Now this requires a transfer of funding from central university administration back to the researcher. That might be an interesting process. At least under UK rules we can now apply for author charges in grants.
Maxine has some good points but can I turn them on their head? If you had to pay referees, let us say the UK research council calculated full economic cost for their time, what does that do to the current business model?
Actually it would be interesting to know what the correlation was between journal impact factor and the relative contribution of personal and library subscriptions (and I guess advertising) is. Does anyone know whether that has been studied?
@Cameron: It is true that journals value peer-reviewers enormously for the work they do. I think most scientists are happy to review for us because they know they will get a similar service when they themselves are authors (as peer-reviewers and authors are in the same scientific pool).
I am surprised to read, Cameron, that you had to pay to reproduce one of your own figures in a review.
I am aware of the argument about taxpayers, having read it previously on more than one occasion. I believe it is oversimplistic to state that direct taxation has a particular, defined relation to a paper as published in the journal, as there are a myriad of confounding factors and steps inbetween, by no means ll connnected to the actual process of generating a paper and publishing it in a journal.
That should be “all” in the penultimate line above.
Cameron, I didn’t properly answer your question about peer-reviewers. In my first comment, I mentioned several other ways in which journals incur costs. And the peer-review process in itself is not a simple matter of publishing papers that reviewers find technically correct: of those, for example, Nature might publish fewer than half, and it is the editors who judge the interest according to a number of factors. Selecting reviewers, also, is not easily transferrable from the journal editorial process to a research council, whose expertise is (necessarily) different (grant funding does not use the same criteria as journal publication).
As Henry said in another forum (Ask the Editor, I think), it would be illuminating for people to come and do internships or undertake other experience in our offices, because I think they would be quite surprised by what is involved in the publishing process. Editors joining our staff (who have usually done a few postdocs in excellent labs and institutions) are invariably surprised, in any event.
Cameron, if you’d like to visit the links I provided to the Nature web focuses in my first comment, you will find articles that work through some of the details you raise in your first paragraph. Perhaps others can provide links to more recent articles here.
Maxine, sorry I wasn’t clear. I paid to reproduce someone else’s figure. Essentially these days everything goes through a commercial copyright clearance centre and so I am not sure whether I would have been charged for my own or not. In practice I would modify the figure from my ‘original’. This obviously depends on the copyright restrictions applied by the journal.
I’m didn’t meant to suggest that refereeing should be transferred to research councils (though its an interesting idea). Just that at the end of the day the publication process is being paid for to some extent by government research/teaching budgets at the moment and that in a full cost model there is an element both of direct payment (subscriptions and page charges) and indirect subsidy (through refereeing and indeed authoring). It seems to me to be more transparent to wrap up the direct element into one author side fee. This gives the researcher control and choice and creates a market where different models can compete.
The issue of papers that don’t get published is obviously an important one. I don’t think anyone would dare suggest submission charges, at least not yet, but maybe it will come. Or perhaps not allowing people to submit unless they have also refereed. Maybe scientific publication is simply not economically viable anymore
Haven’t had a chance to read the web focus you link to but I think you’re point about us not knowing what goes on inside the journal is a very good one. I would love to do it but its not currently high enough up my priority list. One day perhaps?
Bah, my comment above about scientific publishing originally had carefully inserted ‘tongue in cheek’ markup which has obviously been edited out by the system. It was a joke folks!
Publication in PLoS Biology costs $2750. Publication in PLoS ONE costs $1250. Fee reductions are available for Institutional Members. However, in an interesting shift, I have heard through the grapevine that because of the enormous number of submissions published in PLoS ONE (with each one providing the appropriate publication fee), this venture has begun to provide ample resources to not only support itself, but could also begin to partially support other PLoS publications. This is anecdotal, but provides a twist on the author-supported economic model for OA publication.
In retrospect, the high volumes at PLoS ONE make sense; people have a lot of negative data, incremental advances, or general data surveys that they would love to get into the public domain.
Just FYI for those interested: the Journal of Neuroscience (and several other journals) provide an author-supported option to publish OA. This costs $2500 in addition to the regular publication fee of $850.
I also have to agree with Martin that especially in experimental biology, these costs are a small fraction of the total amount of money needed to produce a manuscript, so in most cases (in most Western countries), these publication costs will only hurt a little bit.
The Welcome Trust has commissioned two reports back in 2003 on the economics of scientific publishing: Economic analysis of scientific research publishing and Costs and business models in scientific research publishing. In essence, these reports find that the author-pays model is a viable business model.
Cameron, the Journal of Neuroscience (and probably others) charge a submission fee, in this case $100. A submission fee makes sense if the acceptance rate of the journal is low, because of the costs involved before paper acceptance. The (2nd) Welcome Trust report mentioned above goes into more detail.
@Noah, if that’s true about PLoS ONE its great news. Others of us have published there just to get stuff out quickly as well. Not just because its not wildly exciting data.
“Show me the money?”
At least in UK terms…..
With regards to the Wellcome Trust, the following arrived overnight by email via the American Scientist OA Forum.
For those that don’t know:-
The Wellcome Trust is an independent charity funding research to improve human and animal health. Established in 1936 and with an endowment of around £15 billion, it is the UK’s largest non-governmental source of funds for biomedical research.
Date: Thu, 10 Apr 2008 10:15:52 +0100
From: “Kiley ,Robert”
Subject: Fair use is not enough
At the Wellcome Trust we also believe that “fair use is not enough” if the benefits of text and data-mining – with its promise of discovering new knowledge – are to be fully realised.
Consequently, as a condition of paying an open access fee, the Trust requires publishers to licence these articles such that they may be freely copied, distributed, displayed, performed and modified into derivative works by any user. Publishers may impose conditions on users in relation to attribution (i.e. users must attribute the work in the manner specified by the author or licensor) and commercial use (i.e.specify that the work must not be used for commercial purposes.
All publishers which offer a “Wellcome compliant” OA option – whichincludes, Elsevier, Wiley-Blackwell, Springer, OUP etc – now include this licence information in the XML they deposit in PMC. Some publishers (e.g. Springer, OUP) use the CC-BY-NC, and others (e.g.Elsevier, T&F, Society for Endocrinology) have defined their own licences, but again they explicitly allow text-mining and the creation of derivative works.
These articles are also made available through PMC’s OAI interface, and as such can be downloaded and exposed to text and data-mining services.
Conscious that this licence only extends to “gold” OA articles, the Trust is continuing to work with publishers to explore the possibility of developing a similar licence for author manuscripts.
Regards
Robert Kiley Head of e-Strategy
Wellcome Library
183, Euston Road, London. NW1 2BE
Tel: 020 7611 8338; Fax: 020 7611 8703
mailto:[email protected]
Library Web site: http://library.wellcome.ac.uk
Henry: When I first started using the Nature Network, I posted a formatting blogpost entitled Sweet Sixteen where I go into rather unnecessary detail about such things as changing the colour of text, but it is one of my favourites (of my own) and I was only 16 at the time!
Andrew Sun did a much more elegant post which can be found here
This is so embarrassing. I was 21 at the time. It’s been a long week, and now I have written two comments that are unrelated to the main topic of the blogpost (which I have enjoyed reading, but to which I don’t really have anything to add).
Open Access: Show us the Monkey!
And now, the onion grinder….
Lots of deeking it out online, but FWIW, I’m tuned in here via Windows Media Player. (Real Player kept on switching off every 5 mins).
Take your own pick here
Harold Varmus (Nobel Prize winner, co-founder of the Public Library of Science, former director of the NIH, current president of the Sloan-Kettering Cancer Center, and committed friend of public access) will be on this week’s NPR show, “Science Friday” talking about the NIH
Public Access Policy. If you’d like to tune in, he is scheduled to be on at approximately 3:15 EST.
See here for details.
And if you’d like to call in with a question, you can do so by calling 1-800-989-8255 during the show.
Aagh! I’m drowning in comments! Help! Help!
Thanks, folks, for lots of interesting thought, and links.
Martin and Noah have both commented that costs of publishing may be relatively small compared to other costs. I guess that’s true for some areas, which may then influence which areas have taken up open access most readily. I work in a maths and stats department, where our consumables are actually edible. Or drinkable. Or rather, the water that is filtered through them is drinkable on a good day. So for us $1000 is a lot of money. I suspect it’s the same for ecologists ($1000 could easily pay for a season’s fieldwork). But how many microarrays would that buy?
I was musing about this a bit more, and it struck me that if we had a culture where we were expected to pay for papers, we would do it, and perhaps grumble about the costs. It would be recognised that funding should include money for publication, and universities or institutes would have schemes to help those without project money. So, if we want to have an open access model, the problem is one of changing the culture.
Oh, and on PLoS One being a good place to send negative studies, there is an alternative for people in evolutionary biology and ecology. I felt I had to give it a plug. 🙂
Sorry to add one more, Bob ;-), but just in case anyone is in any doubt about the Nature journals actual policies in this matter, in the light of the plethora of information above, the new NIH announcement makes no difference to our authors, as we already encourage them to self-archive in PMC 6 months after publication, and also allow them to upload preprints into a server such as ArXiv or Nature Precedings, before publication.
Apologies if this reads like a plug, but I thought that in the light of the above, there may be some confusion about our position.
all the best.
There’s a post at Spoonful of Medicine alluding to publishing models. Made me smile.
BTW, when I wrote “30 editors for 20 papers”, I meant editors who handle manuscripts. There are a lot of other editors doing other things (subediting, production, art, News and Views) and other people supporting publication of papers (press office, web production and so on).
This thread is long but I cannot not comment here. I guess I’m one of Maxine’s evangelists and to save anyone checking my profile I’ll declare my interests up front. I worked for Nature for 11 or so years and for the last two and a half I’ve worked for Public Library of Science on PLoS ONE.
Showing you the money is easy. The money for science publishing, whatever the business model, sits in the coffers of those who fund scientific research: research councils, charities, what have you. It also lies in the advertising budgets of companies that advertise in scientific journals but you could argue that since most of those make products bought by researchers then it still comes from the funders of research as a percentage of such companies profits but I digress.
So how does the money get from the science funder to the science publisher? Two main ways:
Either
The university (or department or research group) top slices the funding for their university’s research to run admin some of which goes to support the library which buys access to journals from publishers through subscriptions or site licenses.
And/Or
The researchers themselves pay the publisher costs towards the publication of the papers that they author. Given that publishing the results of research is an integral part of doing that research in the first place more and more funders are explicitly allowing for this to be a part of grant awards.
There are a few more convoluted routes as well such as researchers paying membership fees to a society which then funds a journal or sometimes journals can get charities to support them, though these are often charities which also fund research itself. Anyway however it works chances are the money is coming from the same initial pot of funding.
Personal subscriptions aren’t that common except for a very few journals that manage to incorporate a large amount of ‘magazine’ material. Stuff you want to read in the bath after a hard day pipetting. Personal subs to most journals are from lab heads providing journals that they see as being needed for their work, an expense of running a lab that may well be tax deductible.
You may have noticed that I haven’t mentioned Open Access yet. That’s because all of the above is about the business models of publishing and Open Access isn’t a business model. Open Access tells you what you get for the money, however it is supplied. I think that Open Access provides a more useful primary literature by reducing the barriers that the conventional publishing puts in the way of the reading and use of that literature. Perhaps that’s just me; perhaps I’m naïvely idealistic.
I don’t want to prolong this comment too much, so I won’t talk about the problems of even relatively well supported researchers in ‘developed’ countries getting access to all the literature they might need, I won’t talk about the problems of text mining and machine reading of a closed access literature, I won’t get into the ethics of copyright and how it should be an enabler of information sharing not a bar to it. There will be other times and places for that.
So my answer to the challenge of this thread is a retort. Don’t show me the money, show me the results!
Chris – that’s not a comment, that’s a blog post. Nu?
Yeah, I know and I apologise. I must get myself a blog so I can rant on my own dollar.
I didn’t mean to sound condemnatory – far from it. It’s just that you had a lot of things to say that were important for people to read, and posting them in a blog would make these things easier to read and to get at than having them buried at the bottom of an unlit staircase behind a door marked ‘no girrafes on unicycles beyond this point’.
You might find it difficult to believe, Henry, but this isn’t your blog.
It’s Bob’s. And I kind of imagine it in a glass-fronted high-rise with smart, uniformed attendants, with lush plants in the foyer and Bach playing softly in the background.
Crikey. I didn’t mean to dig a hole for myself, let alone make it any deeper. It’s just that – speaking for me, anyway – I get a lot of ideas for blogs from comments I leave elsewhere. So I leave a shorter comment and expand the argument into a blog post. Phew. And I hadn’t imagined a locale for Bob’s blog, but whatever it is, it has a cat in it.
Henry, you’re …
squirming.
The attendants are told not to bother with being smart, because they’ll only get cat hair all over them. It’s really not worth the effort – we all know who’s in charge.
Chris’ comment isn’t that long – I’ve seen much worse. The only thing I would add is that open access doesn’t work with buying subscriptions, which is where there is a connection with the business model.
Oh yes, just so there’s no doubt or confusion on this point – unicycling girrafes are always welcome.
Well you could have subscriptions to an Open Access journal, especially if what you got for your subscription was a mailed hardcopy or some other subscription advantage, kind of like The Economist. It is probably not going to be a sustainable business model for most science journals though if that is the only source of revenue.
Then again I have a strong suspicion that if Nature made all its primary research articles Open Access, while all the other content (News, N&V, Reviews, etc) was kept as a subscription advantage, they wouldn’t see too much of a drop off in subscription revenue. Meanwhile the kudos they would get for doing this would attract even more of the best papers and so increase their other sources of revenue.
Oh dear. I feel another GRM (Golden-Retriever Moment) is due.
