Thursday, 19 July 2012

The Open Access Publishing Revolution

After a campaign championed by the Guardian, the UK research councils have announced a move towards open-access publishing. There's currently a huge amount of debate within the academic community about the best model for publishing.

Currently, academic publishing is dominated by the large publishing houses - Elsevier, Wiley-Blackwell, Springer, etc. When I write a paper with new and interesting scientific findings, it needs to be published to enable my fellow scientists to read about what I've done, enabling them to use my findings in their research. The typical business model for a journal is as follows: most journals do not charge me a fee to publish (some do, but it's usually not too onerous). They make their money by charging people to read them. University libraries will buy subscriptions to the journals (usually they will buy 'bundles' allowing subscriptions to an entire publisher's catalogue), allowing all academics at the university to read the article. Meanwhile, a company wanting to access the research is more likely to pay the upfront fee for an individual article, which is usually about £30. Similarly, private individuals wanting to access the research will have to pay the £30 fee per article.

The dissatisfaction with the current model stems from the fact that the majority of scientific research in this country is funded by the tax payer (including my own). How, then, can it be fair that tax-payers are paying once to fund the research (out of their tax money), and then have to pay again as individuals to read the research? Many in the academic community view this as a morally untenable position - the public should have free and unfettered access to the results of all taxpayer-funded research (and when you read the comments sections on various websites, you really do get a sense of moral outrage).

I disagree with this argument. There are many things that my tax-money pays for, yet I can't just access for free. Our taxes go to the BBC, but I still have to buy a license fee to watch it! Taxes fund the NHS, but I still have to pay a fee for my prescriptions. Let's face it, some of my tax money goes to the army to buy tanks - that doesn't mean I have a right to turn up at the gates of my nearest barracks and demand a ride on a Cheiftain!

But, under the pressure of the academic community, the government, via the UK funding bodies, has buckled to pressure and is now about to make things far worse. They are planning to mandate that we publish only in open access journals. The business model here is that, rather than a 'reader-pays' model, we switch to an 'author-pays' model. Journal articles are made available for free, so instead authors are charged a much larger amount (typically £2000 or so). As far as I am concerned, this is a very bad publishing model!

What are the benefits of a so called 'gold' open access publishing model, where the author pays? Well, the sole benefit is that anyone, anywhere, can read your article for free. The negative is that this could add an extra 5 to 10% to the costs of of UK research. For example, last year NERC (my funding body) spend £180 million on research, and about 5,000 papers we published as a result. With gold open access of £2000 per article, this would be an extra £10 million, so just over 5%. Of course, the funding councils are not being given any extra money, so that means they will be able to fund 5 to 10% less research. In a time where our science budgets are increasingly stretched, and grant money harder and harder to come by, can this really be a good thing?

Secondly, we must consider the pressures this will place on the publishing industry. Currently, in a 'reader pays' model it is in the publisher's interest to publish only the best and most relevant research. A journal cannot afford to waste money publishing work that will not get read. Therefore, the commercial pressure is on journal editors to accept only the best work. Under an 'author pays' system, the only commercial pressure on the journal is to publish as much as possible. Quality is no longer a driving factor, because it doesn't matter whether things get read, all that matters is that there's lots of papers. So an author-pays model will simply lead to a significant reduction in the quality of articles that journals are prepared to accept. Putting this as simply as possible, there would be no reason for a journal to ever reject any paper, ever. I don't want to spend my time wading through a ton of crap papers to find the one or two good ones that I need. I want a journal to have already taken editorial decisions to only bring the ones that are of the highest quality to my attention. 

What about the effects on libraries? Perhaps the libraries would be able to save money, because they wouldn't need to pay for so many journal subscriptions, and the money saved could be transferred over to cover author charges. However, this wouldn't cover all the back issues to which scientists need to access. Nor would it remove the need to subscribe to international journals to access work from people in other countries. Science is an international effort, so if the UK does something in isolation, it won't affect the need to pay subscription fees to read about work from every other country in the world.

Finally, what will be the impact on publishing academics. Obviously, the only barrier to publication will be money, so if you have money, you can publish, while if you don't, you can't. The model in mind is that university faculties will have central pots of money to pay for open access publication. So who will get the money for publication - the junior PhD student who has made a cool new finding, or the senior professor with his hand on the purse strings? At present, the quality of the paper is the sole deciding factor in where a paper is published. Under the proposed system, the deciding factor will be money. This would NOT be a fairer system, if would be significantly UNFAIR! I cannot understand how academics can be in favour of such a system.

The other proposed model is so called 'green' open access. This model takes the money out of the system entirely. Academics post their own content on their own websites (or possibly on a university-wide archive) or on archive sites such as the arXiv. To a certain extent, this system is already active, because many authors do post their work onto arXiv, and do post their papers on their own websites. Many academics see this as the model to which we should evolve. However, I think this is somewhat misguided, because the currently green open access exists off of the back of the mainstream publication industry, it cannot exist without it. Firstly, note that many green open access sites are heavily subsidised (as arXiv is by Cornell, for example, to the tune of about $500,000 per year).

More importantly, if I publish a paper in arXiv, it is not peer reviewed. Peer review as a process is vital to science, it makes sure only reliable results are published, and removes all the dross. I know it's not perfect, but it's the best system we have. The only way to get things peer reviewed is to submit them to journals, where they will undergo the full editorial process. Currently, journals are happy to provide this service, and publish the paper as their content, while allowing re-prints to be republished on author's websites, because it's not done widely enough to effect their margins. Just because a certain number of authors have put their papers on other websites, libraries are still forking out for subscriptions. If green open access became so widespread that it began to impact on the bottom line, I expect journals would begin to clamp down.

I guess this all comes down to whether publishers provide a useful service, or can we do without them. If they provide a necessary service, we should be prepared to pay for that service accordingly. If they do not, or if we believe we can achieve the same effects more cheaply, we should attempt to do so. Paraphrased from a comment on this blog, we need journals to provide (1) a decent and enforced peer review system, (2) an editorial system which can reject (or grade) papers (which provides quality control and therefore acts as a marketing tool) (3) a publishing system and (4) a citing system. The need for peer review is obvious, as is the need for a publishing system. Both cost money to administer. It is not free to put things on the internet - arXiv needs half a million dollars a year to function.

The need for a citing system is a key problem for green open access. Say I publish my paper, and put it on my website. Will it still be there in 15 years time when I have long moved on, or will a searcher simply find a broken link? It's vital that researchers can access older papers - journals provide the means of archiving them and an easy means to index and access them. This would not be possible with authors sticking their papers wherever under a green open-access system. The editorial and grading system is also vital, for both author and reader. Rightly or wrongly, as academics we are graded on the quality of papers we produce. When we can post whatever we like in an open-access mega-journal, it becomes very difficult to assess the quality of a academic's output (perhaps this is why so many favour this system). More importantly as a reader of papers, I much prefer to be able to look at the monthly output of my favourite journals, scroll through the list of papers that they have (which is a manageable list), rather than dive into a mega-journal and have to sift through all the dross to find some good papers. I'd rather pay an editor to do that for me.

Given these needs, how much will all these things cost? Let's imagine a typical medium-sized journal. They'll probably have about 5 staff: a senior editor to lead the editorial direction of the journal; two publishers to facilitate the peer review process - selecting the reviewers for each paper (or rejecting the really rubbish papers without review; an admin; and a guy to do the publication side of things. Plus within the publishing house they have the IT support and the rest. Let's say it costs £500,000 a year to cover everyone's salary, pension, computing costs, web archiving costs and the rest. This is a complete guess, but it doesn't seem like an exorbitant amount, right? In fact, this is probably on the low side. Well, for a medium-size journal publishing 500 papers a year, that's a cost per paper of £1000. So this must be paid either by the author, or through subscription. I don't see green open access as long term sustainable except on the back of a healthy publishing industry to bear most of the costs. As a commentor on the Guardian website put things rather well of the problems we'd face if we abandoned journals:
The real problem with getting rid of journals is that it means that there is no chance of anyone reading your work if you are a junior research at a non-high tier institution, which means there is very little chance of getting any recognition or promotion.
If you are already a big name in your field, or you are a young researcher at a place like Harvard/MIT then you could wipe your bum on a bit of paper and put a photo of it on your website, and people would download it to look. But if you are a postdoc at a place outside the top 50 (or so) then noone is going to read anything you write unless it is published in a good journal; does anyone really think that academics are regularly checking the website of Podunk State University to see what their guys are doing?
The main role of the journal system is to filter papers based on their quality, so that the ones that are worth reading on average end up in good journals, while the ones that arent worth reading end up either unpublished or published in so low a venue that noone willl find them. This is a good system because it means that academics can just follow a handful of general interest journals along with the websites of those they know are good in their field, without having to wade through the tens of thousands bad papers that are produced every year. It also means that any academic will get their work widely read as long as they can get it into a decent journal; which would absolutely not be the case if we abolished journals.
So, we come back to the question at hand, if we have to pay for a quality publishing industry, who should pay? For the reasons listed above, I think it should be the reader, not the author.

My thoughts on this are formed partly by my experiences while working as an intern at Shell. During my 3 months there, I must have downloaded at least 20-30 papers, paying the upfront costs each time. That comes to about £600-£900. Shell have a library system where you request a paper and they go off and get it for you, and they don't even (appear to) worry about the costs. I've no idea how much the average Shell employee downloads papers, but these are science-intensive companies, so I bet it's quite a lot. Whenever I meet employees, they all seem fully up-to-date with the latest research, so they must be reading it. This is money that goes back into the academic system, helping to contribute to a healthy publishing system that is necessary for research. And Shell are going on to use this research to make money. Under an author-pays model, they wouldn't have to contribute anything.

So what about the poor UK 'citizen-scientists', working independently, who want to read the latest research papers. These people are usually all over the Guardian comment pages complaining that they can't access research without paying. My first opinion is that they can't be that keen to read it - if you want to read a paper but can't find a non-paywall copy, your first port of call should be to email the author, who will likely be more than happy to furnish you with a copy. If you can't be bothered emailing the author, then frankly you can bugger off and stop moaning. Secondly, accept that the changes you require to the academic publishing industry would be seismic, causing immense disruption. I'm not allowed to pop over to my nearest air-force base to have a go on a Eurofighter. It would be immensely disruptive, even though it was paid for with my tax. Similarly, you don't have the right to cause immense disruption to academics just to fuel your hobby.

However, I think there should be a much more simple solution. We should develop a system where independent people are capable of registering with their local university library. Perhaps for a small admin fee, this would give them an online library log-in, giving access to all the journals they want to read. I can't imagine this would be taken up by more than a tiny fraction of the population anyway. This seems like a far more simple system to set up, rather than insisting that we either destroy the publishing industry or spend 10% of the UK's science budget on publishing fees that we have no need to pay.

And if you have a problem with Elsevier coining it on the back of your research, don't publish with them. There are loads of journals that are run by various learned bodies (in my field, for example, GJI is published by the RAS, GRL by AGU, Geophysical Prospecting by EAGE, ERL by IoP). Any money made by these journals is pumped back into the subject by the learned bodies, funding conferences, research fellowships and the rest. So if you don't like the big publishers, find a journal run by a subject and publish in there instead.


  1. Chieftain? Chieftain? I know we're children of the Cold War JV, but there's no need to go back to the 80's yet. Challenger 2 know don'tchaknow :P

    1. My tank knowledge is, I'm afraid, sadly lacking. I propose open access to all tax-payer funded tanks!

  2. What do you think about the peer review through the wiki-method? I don't know much about it, but I understand it's becoming more common in physics.

    1. It's not something I've come across. Any idea where I could pick up some more details?

    2. I'll have to rummage. I'll try Tom Chivers blog and New Scientist first, I think I saw an article about it on one of those. Basically the idea was to have an academic papers wiki, and then the peer review comes from everybody who's on the site, allowing people to judge the quality of the paper from who's reviewing it and what their opinions are.

    3. Is the idea that people can leave comments, or that they can actually edit (as per wikipedia)? I'd by extremely uncomfortable with a system where my work could be edited by others where I had no control. After all, if it's my name on the top of a paper, it's my reputation that's on the line.

      I see no reason why allowing comments on a paper would be a problem. However, I'm not sure it would be as effective as peer review as it currently stands. Firstly, as reading just about any online comments section on any platform anywhere would reveal, is that without editorial control you'd run the risk of degenerating into a slanging match, or rapidly disappearing at a tangent over an pedantic points. I suspect you'd also have many mediocre papers with no comments or review whatsoever. While you might blame the author for writing something that simply isn't interesting enough to attract an online commentary, this work still might be worthy science deserving of peer review.

      So while these might be good things to try, I see them as being supplements to a strong peer-review and editorial process, rather than a replacement.

    4. Right, web filters are down, and I can blog again. I can't find the original article, but the principal is called open post-publication peer review, and a blog explaining much better than I can is here:

      So that is only allowing comments, which the author can then respond too, and I imagine edit his own work to tweak post-publication (a sort of ongoing errata, I guess). So not actually a wiki at all, but you could have the author able to edit his work, and then editors (administrators), who can then apply the usual rules of banning anyone who doesn't play nicely, and you can gain a reputation by contirbuting reviews, and so work your way up to being an editor. Disappearing into tangents can be spun off into a new thread. The incentive for people reviewing mediocre papers would then be to gain experience and reputation. Reviews can be made anonymously, as with the current system, or openly, to show participation in the process, and gain reputation. Well, I think that's the idea anyway. I'm just glad to have an internet connection for a bit.

      Of course, all that means that people are still using their time to review stuff, so it depends which is cheaper, post-docs time, or research funding? Having it all electronic would save money mind.

      There are resarch systems whereby you can gain a reputation and be able to edit stuff, but seems to be confined to maths at the moment, where I guess it's easier to be collabarative. An example would be MathOverflow which is for maths research questions, so it's more to do with collabarative research rather than publishing.

      Dibs on the first open-access tank ride.

    5. Hmmm, interesting idea - I guess in a few years we'll see how effective it is. My first consideration would be that such a forum would still require a fair amount of administration and moderation, which means employing people, so I'm not sure how you make publishing cheaper by doing so.

      It may work for us geological types where our papers are not of immediate significance, but I think this method would cause trouble in the Life Sciences sector, where life-affecting medical treatments are decided based on peer-reviewed literature. One could easily imagine a person with a crank treatment (homeopathy, say) submitting a paper and getting a bunch of mates to swamp it with positive reviews. This would give the paper an imprimatur of peer-reviewed decency which it wouldn't deserve.

      Actually, I say this wouldn't be an issue for us geo-types, but actually I could easily see this happening with fracking! For any controversial scientific topic, any topic with any social or political ramifications, this method would be a nightmare (perhaps that's why it works for Maths). The point of the editorial review process is to find reviewers who are sufficiently qualified and likely to be impartial (I know it doesn't always work that way, but that's the idea). It's shouldn't be a free-for-all popularity contest.

      One more thing - I don't think academics will ever improve their reputations by reviewing work. Doing reviews is seen as a bit of a bother but something we have to do to give something back to the field. If we can get away without doing it, I think we would like to. Unless your reviews count towards your REF score, we're not going to care about doing reviews to boost a reputation.

  3. REF Score?

    You raise good points with issues like political and controversial subjects. Though that could also happen today, by those people publishing their own journal, and then referencing each other's work. And yes, as the scale of this increased you would need people who would act as editors and moderators full time. Perhaps that would be cheaper than the current proposals? Can't say I see a way to tell.

    1. REF = Reaearch Excellence Framework. Actually, it's now called the RAE, Research Assessment Exercise. It's how we're assessed as academics - they look at all your papers and give you a score. The better scores a dept gets, the higher it is rated, and this can affect how of the pie a dept gets. So depts try to hire people likely to get good rae score. So it has a direct impact on our careers. So unless reviewing is made to count to the Rae, it won't be our priority.

      You're right, cracks can set up their own journals, but in that system it's pretty easy to spot the bollock journal of made up woo and ignore it. If everything is in together and you have to rely on reading all the reviews, it'll be harder to discriminate.

      But overall, the key is that we don't know. Which worries me that we're taking career-affecting, multimillion pound decisions based on a bit of a hunch, some guesswork and a general academic unease of all thing corporate.

  4. Maybe you could make reviewing count towards RAE. If it means you can get a bigger part of the pie, and more prestige, great! Reward people for contributing to the common good.

    To clarify one thing, I wasn't saying you'd have to have only one site to go to publish articles. I'd imagine things could self-organise into different specialisation areas quite easily. It would be useful to have a registry of everything on all the different site, for search purposes., but there'd be nothing to stop people collaborating to make a site for publishing on a particular subject, with their own moderators and administrators, which could then function a bit like a journal does now. It's just a matter of who you rope in. Sorry that's a bit hand wavy, just doing this in my lunch hour.

    1. Reviewing is easy compared to doing new science, so to make it fair it would count a tiny fraction of an amount towards a RAE score in comparison to having a catalogue of good papers.

      Obviously you'd have different sites for different subject areas. But cranks (and non-cracks who have the misfortune of writing crap papers) in a given area still want to publish in the same journal as everyone else. Without editorial selection they'll all end up in the same place, and the reader will have to sift through. As you say, you can have moderators and administrators to help sort things out, but this would probably need to be a full-time job, meaning someone needs to get paid, needs an office, needs a pension scheme and health insurance, which all costs money, meaning the journal has to develop a revenue stream from somewhere, which leaves us back where we started.

  5. So journals or journal-like entities (lets call them journalomorphs), evolve due the selective pressures of publication credibility. Interesting. Maybe someone could put together a research funding proposal, where they can run a computer simulation model where academics compete for funding and grad-students by gaining prestige and positive paper-reviews, and see if journalomorphs naturally form in that environment. Maybe give us clues for journal structures that haven't even been thought of yet. META-RESEARCH!!