Saturday, 26 February 2011

Academic publishing: why isn't psychology like physics?


My job is to conduct research and publish the results, so I’ve become interested in discussions on the internet about models of publishing. There are three tensions in the field: (1) who pays to publish research? (2) who decides what gets published? and (3) who takes any profits?
In the traditional model, based on publishing on paper, the answers have been (1) readers pay, via journal subscriptions, (2) editors decide, advised by reviewers, (3) publishers take the profits.
But publishing on paper is in decline. Most of us read papers on the internet taken from electronic versions of journals; either directly on computer, or printed from a pdf. This has led to some re-evaluation of how publishing is done. Most publishers, with an eye to their profits, have attempted to just continue with the old system, charging subscriptions for electronic journals much as they have done for paper journals. These are largely paid for by institutional libraries rather than individuals. If your institution doesn’t subscribe, or you don’t belong to an institution, there is an option to pay for individual articles, but the charges are typically prohibitively high.
Psychologist Stevan Harnad was among the first to point out that the internet changes the publishing game completely. Way back in 1994, he suggested that academics could leave publishers out of the research communication cycle altogether. For instance, when I write an article, I could post it on my website, and anyone who wanted to read it could download it. Instead of institutions paying thousands of pounds in journal subscriptions, and individual readers being clobbered for $30 for a five-page article, academic exchange would be virtually free. What would be wrong with that? This appealing idea has over the past sixteen years fostered a huge amount of debate. What is surprising to me, though, is that there appears to be a massive discrepancy between disciplines in how far Harnad's ideas have been taken up. And Harnad's own discipline, psychology, is one that has been very slow to respond, compared, say, to physics - see Because many of my readers are psychologists, and many psychologists are wedded to the traditional printed journal, I'll briefly rehearse some of the arguments that have been made, before commenting on the possibilities opened up by e-publishing.
There are two obvious snags to bypassing journals. The first just concerns presentation. Traditionally, publishers have ensured that journal articles look nice: properly formatted, intelligible figures, grammatical writing, and so on. Their role in this aspect of publishing has, however, dwindled over the years. This does not mean that any academic can produce a nicely-formatted article: some skill is required. And many academics don’t write very well and benefit from the services of professional editors. But increasingly, publishers are cutting down on professional editors and graphic designers - indeed, increasingly they hold authors responsible for getting their figures formatted for direct web publication (see my post).
The second point has to do with quality control. Traditionally, journals have acted as gatekeepers. If an article appears in a peer-reviewed journal, it can usually be assumed that it doesn’t have serious flaws: the higher the quality of the journal, the safer this assumption should be. And for the top-notch journals with the highest impacts, publication of paper generally gives some guarantee that it is not only methodologically sound but also particularly important in terms of advancing the field.
Nevertheless, some have queried whether we need this traditional function of publishers, pointing out flaws in the peer review system and idiosyncratic behaviour by editors. Indeed, gatekeeping may not just be imperfect: In some fields it may damage science, by inducing a publication bias in favour of ‘significant’ (both statistically and theoretically) findings, and biasing against publication of replications (or non-replications).It's even been argued that publishing in high-impact journals may be unethical, because such journals are particularly likely to publish papers that over-state their conclusions. And unless there is a massive media stink, they resist accepting papers that fail to replicate exciting findings that they've previously published.
Such considerations have led some to argue that we should ditch not just publishers, but the whole editorial and reviewing process as well. In effect, everyone would just publish their own stuff, and it would be judged by the academic community. Although this solution appeals to the more anarchic side of my nature, I can see objections to it. The main is that we’d all be swamped by a tidal wave of information. The good stuff would be buried within a heap of rubbish. And because methodological flaws are often hard to detect unless you have done research in the field yourself, problematic but apparently important studies might have a large impact.
As Harnad noted, however, taking publishers out of the loop need not entail abandoning peer review. With the exception of a few top-notch journals, most academic publishers don’t employ editors. Instead, editors do the work for free, or for a small honorarium. It is very unusual for reviewers to be paid anything. So this gate-keeping role of publishers doesn’t need to reside with publishers. Any group of academics could decide to set up a journal, appoint editors and operate a reviewing process. Provided the journal was electronic, with no paper copies, this should be achievable at relatively modest cost.
But has it worked? Well, sadly, not for psychology. I've brooded about this issue for some years, but was stimulated to revisit it by my recent discovery of Kindle direct publishing. It’s different from an Open Access model, because the reader pays for content, but the sums involved are trivially small compared to current journal charges. The Kindle operation is designed for publishing books. Harnad has always distinguished commercial book publishing from what he terms ‘esoteric’ publishing, i.e. publication of academic articles of interest only to a small group of experts. But my Kindle experience made me wonder whether the distinction between the two types of publishing need be as sharp.
I came across Kindle e-publishing when I’d had my first novel turned down by two agents, both of whom spoke of a ‘saturated’ market for crime novels, the genre I’d written in. I’m not a typical author: though I’d be very pleased if people enjoyed reading my book as much as I enjoyed writing it, I don’t have any high expectations of a successful second career as a novelist. I write just for fun, and, being an academic, I’m used to not getting paid for what I write. And being an academic also makes me used to rejection, while at the same time recognising that I shouldn’t let it discourage me if I have a belief my work is good.
After my experience with open access journals, I’d expected that any business that offered a self-publishing option would charge authors, especially for ‘esoteric’ material. As a first novel by an unknown author and with neither agent nor publisher, my book was definitely in that category. So I was pleasantly surprised to find that not only did I not have to pay anything, but I could set my own price and get a 70 per cent royalty. For a book with low-volume sales, this is going to be negligible, especially as a great chunk of tax is taken off any US sales. (This can be avoided if you fill in a lot of forms, send away your passport to the US, and so on, but since I was not expecting more than about two US sales, this clearly was not a sensible option). Anyhow, you are walked through the process with easy instructions, which include downloading a couple of bits of software to convert your manuscript into the right format, and then you press a button and upload your book. The whole process took about the same amount of time than it typically takes to submit a journal article through an electronic portal. And  you don't need a Kindle, though the reading experience is better than on a PC or iPhone. I keep wondering if there’s a catch; no doubt someone will tell me if there is. In common with some others, I am a bit dubious about accuracy of the sales figures provided (there's a rather surreal exchange about this on the Kindle forum). But when I go to Kindle store, there it is: the Case of the Fremantle Fingers by Deevy Bishop at $2.99. This gives me a small glow of satisfaction.
So I started to think, couldn’t research papers adopt this model, but with even lower charges? I’m not saying the articles need be in Kindle format: I don’t think it’s ideal for scientific papers. But rather that the method of author download to a central site, from which others download at a minimal cost, would be an interesting one for psychology (and other discipline) journals to adopt. The difference from current practice is that the material would be affordable; instead of costing 20 to 30 dollars, individual articles would be priced at 20 to 30 cents. The Kindle store does sell some books, often out-of-print classics, for ridiculously low prices, (or even free), and the charging mechanism doesn’t seem to be a problem. Authors of academic articles don’t expect royalties, so these could be waived.
But what about the dangers of the free-for-all and lack of peer review to identify quality? I suspect that even a notional charge per download might be a help here, by identifying material that people would be willing to pay for. And the Amazon system also suggests a more explicit way in which refereeing be adapted to such a system, by using a star rating with optional comment. This is especially useful in the academic context if the reviewers themselves also have ratings. For instance, a 5 star reviewer would be one who’d published significant papers in the same area, a 3-star reviewer would be someone familiar with this or a related area but without significant publications in that area, and those with no expertise would be 1-star reviewers. It shouldn’t then be difficult for readers to see which papers had high ratings from knowledgeable people.
But would people make ratings? There has been disappointing take-up of Comments options for journals that offer this, such as the PLOS series. I don’t like anonymity, but  to encourage frank comments on papers,  there would have to be some way of registering reviewers that made it possible to check their credentials and assign them an anonymous ID. Otherwise, junior people might be scared to make adverse comments on senior figures.
The downside? For me, omitting the usual pre-publication reviewing stage would have disadvantages. As I’ve noted before, I find peer review painful but often helpful in improving papers. Sometimes it has saved me from revealing an embarrassing amount of ignorance. I’d be happier with a system that allowed an option for pre-publication review, and it would be interesting then to see whether papers that were identified as having undergone such review would attract better post-publication ratings than other papers. Learned societies, for instance, which currently produce paper journals, might instead use their existing reviewing systems for papers which are then deposited on the electronic store, with a quality kitemark.
Another drawback is that for some people, particularly those in low-income countries, even a nominal charge per download would be a deterrent. But it should be feasible to devise a system that subsidised those from resource-poor countries. Or  the charge could be waived once a certain number of downloads had been achieved, so that the most popular material becomes automatically freely available.
Since Harnad’s ground-breaking formulation of the Subversive Proposal much has been written about models of science publishing. Many changes have occurred, with increasing pressure for open access, and growing criticism of the peer review system. But there remains a massive difference between disciplines: physics embraced self-archiving rapidly, whereas scientists in other disciplines often don't even realise it's an option. Maybe this has to do with the pace of change and degree of competition in a field: physicists don't want to wait months before their work is published because it could be scooped. Psychology experiments are seldom so time-sensitive. Nevertheless, I think the pace of research in our field could be improved enormously if we broke free from the stranglehold of the traditional commercial publishers. 

btw, Tom Webb has drawn attention to a Commons Select Committee that is seeking evidence on the peer review process. This strikes many people, including me, as a bit strange, in that it would seem to fall outside the remit of government. If you want to comment, you need to do so by Thurs 10th March, 2011.