What should society journals do about open access?
July 7, 2012
I got an interesting email a couple of days ago from Robin Wilson:
Something which I thought may be of interest to a number of people who are keen on OA is your views on what learned societies should do about the closed journals they currently run. I’m particularly interested in this as I have just joined the Publications Committee of a (smallish) learned society in the UK who publish a journal through Taylor and Francis which is currently closed access.
In some ways I want to go to the next meeting and suggest that they make it Green OA – but I can’t see that happening because it brings in a significant revenue stream for the society, thus allowing them to charge far less for their annual conference, employ admin staff, publish a quarterly newsletter etc. Gold OA is possible, but would restrict publication to those whose departments or funding bodies would pay the fee – and as this journal isn’t right at the top of its field the argument may go “Why publish in X for $3000 when I could (try to) publish in Y for free!
It’s an important question. Before I try to answer it, let me make it clear that I have no qualifications whatsoever to comment on this subject. I am an extremely junior researcher — not even a postdoc, I have an honorary position. I have never been an editor of any journal, nor on any publication committee. Regarding the flagship journal of our discipline, the Journal of Vertebrate Paleontology, I am not even a member of the society that produces it, having not renewed my SVP membership after the society’s cowardly abdication of responsibility over Aetogate. So I have no authority, no experience, and no influence. Anyone who is interested in my opinion anyway is welcome to read on. The rest of you should read this comic instead.
Pros and cons
First, there are vcry good reasons why society journals should go open access. Here are some:
- It’s in line with societies’ missions. For example the SVP says “The object of the society is to advance the science of vertebrate paleontology and to serve the common interests … of all persons concerned with the history, evolution, comparative anatomy, and taxonomy of vertebrate animals”. Without question, freeing the research in its journal would be a huge step towards fufilling that mission.
- There is a significant citation advantage. This graph shows the changing impact factor[1] of Acta Veterinaria Scandanavica through time. It’s on a definite downward trajectory from 2000 till 2006; then it goes open-access with BMC, and immediately switches to a strong upward trajectory.
- Increasingly, authors are reluctant to give their best work non-open journals. (For myself, I no longer give any of my work to non-open journals; I realise that not everyone is in a position to make such a blanket commitment, but many people make it at least part of their decision process.)
On the other hand, there are also reasons why society journals might want to remain subscription-only. Here are a couple:
- As Robin noted, the big one is subscription revenue. From a quick browse of the SVP web-site, I can’t find any accounts (which is a bit of a transparency fail), but I am guessing that JVP subscriptions are a pretty big part of the society’s income. That income might be threatened by going open-access.
- A journal that switches to Gold OA (i.e. the author or his institution pays a publication fee) may find that it gets fewer submissions, as authors switch to free-to-submit venues.
(I think I have all the major pros and cons here — did I miss any?)
Evidence
We know from many independent studies — not just anecdotes like the graph above — that the open-access citation advantage is real. A good summary is found in Swan (2010), which surveys 31 studies of the OACA, showing that 27 of them found an advantage of between 45% are 600%. I did a rough-and-ready calculation on the final table of that report, averaging the citation advantages given for each of ten academic fields (using the midpoints of ranges when given), and found that on average open-access articles are cited 176% more often — that is, 2.76 times as often — as non-open.
By contrast, we have no actual evidence for either of the negatives. To be clear, I’m not saying that the downsides are necessarily not true, only that we have to rely on our intuition about them rather than hard facts. We don’t actually know what effect a switch to Gold OA would have on submissions rates, because there are no studies that analyse this. (At least, none that I know of — can anyone enlighten me?) The example of Acta Veterinaria Scandanavica doesn’t help us here: its impact-factor increase says nothing about how many articles it published, only about how widely cited they were.
Similarly, no study has ever shown that Green OA — allowing or even encouraging authors to deposit preprints in repositories — harms subscription revenue. In fact there is good evidence that Green OA either does not affect subscription revenues or, surprisingly, actually increases them by acting as an advertisement for the “official” versions of papers. For example, Swan (2005) wrote:
We asked the American Physical Society (APS) and the Institute of Physics Publishing Ltd (IOPP) what their experiences have been over the 14 years that arXiv has been in existence. How many subscriptions have been lost as a result of arXiv? Both societies said they could not identify any losses of subscriptions for this reason and that they do not view arXiv as a threat to their business (rather the opposite — this in fact the APS helped establish an arXiv mirror site at the Brookhaven National Laboratory).
For much more on this issue, see the PEER project (Publishing and the Ecology of European Research). For example, the project’s final report says that:
A Randomised Controlled Trial indicates that making preprints visible in PEER repositories is associated with more traffic to the publisher sites at the aggregate level, but this varies by publisher and subject. Overall, PEER is associated with a significant, if relatively modest, increase in publisher downloads, in the confidence range 7.5% to 15.5%.
The likely mechanism is that PEER offers high quality metadata, allows a wider range of search engine robots to index its content than the typical publisher, and thus helps to raise the digital visibility of scholarly content.
Gold, Green, subscription or hybrid?
So what should subscription-based society journals do? Stay as they are? Switch to the author-pays Gold OA model that has served Acta Veterinaria Scandanavica so well? Support Green OA by encouraging preprint deposit? Or become hybrid journals (as JVP is now) by offering optional open access within otherwise subscription-based journal?
It will come as no surprise that I think subscription-only is a disaster, and that some kind of change away from that is necessary. A scholarly society simply can’t best serve its discipline by locking its work behind a paywall. It just can’t.
I also have an increasing sense that “hybrid OA” (i.e. a subscription journal with an optional open-access fee) doesn’t really work. Certainly Elsevier have had astonishingly low uptake, and there are good reasons for this. I’ve heard that JVP‘s optional-OA uptake has been disappointing, too — probably the true reason for the recent price-cut. My guess is that this is pretty representative. So I think that hybrid is really a bit of a fig-leaf that’s used by publishers and journals that don’t really want to do OA but feel they have to be seen to be doing something.
So journals that do want to move to open access have to choose between Gold and Green. My best judgement is that Green is easier in the short term, but that most journals will want to go Gold in the end. The advantage of Green at this point is that it doesn’t require the societies to actually do anything — just to tell authors that they are welcome to self-archive preprints, after revising for reviewer comments, but before copy-editing (if any) and typesetting. A forward-looking society might host an archive for this purpose, but that is not necessary.
The thing is, a society that is serious about open access — that really wants that 176% citation advantage — will need to not just grudgingly allow Green OA, but actively shout about it. That’s maybe the single most important point here. Open access is not a threat to be deflected, it’s an opportunity to be grasped. The journals, and societies, that do that most effectively will be the ones that flourish. Right now, JVP is dead to me, and to some few other palaeontologists I know. I’d love that to change.
(Going forward, I am not sure that Green will remain viable indefinitely, but that is a subject for another post.)
References
- Swan, Alma (2005) Open access self-archiving: An Introduction.
- Swan, Alma (2010) The Open Access citation advantage: Studies and results to date.
Footnotes
[1] Yes, yes, I know.
Sauropod Vertebra Picture of the Week 
July 7, 2012 at 3:34 pm
Glad to see you treating this issue. I think that it’s one of the real issues with the switch to open access.
You can see in the SMBE society’s reports that its revenue is mostly from the subscription model journal MBE: http://www.smbe.org/about/reports/
We were told at the most recent congress that the OA GBE now also makes a benefit for the society, but I don’t have numbers.
July 7, 2012 at 4:15 pm
It took societies a little while to start experimenting with open access, so Acta Veterinaria Scandinavica was the first journal in BioMed Central’s stable to convert. However with last week’s release of the 2011 JCR there are more examples of excellent increases in impact when subscription journals move to an open access model. Genetics Selection Evolution is a good example. Owned by INRA, it moved to us in 2008 and is now the top of its category in the JCR. Additionally all the impact factor journals which have moved to BioMed Central and an open access model, have not only seen no drop in submissions, they have generally seen an increase which has allowed them to be more selective if they wish and increase the quality of their publications. We are haopy to provide more case studies for anyone who wants them!
July 7, 2012 at 4:37 pm
Thanks, Deborah, that’s very useful to know. Would be great to have a link to more detailed stats.
July 7, 2012 at 9:10 pm
We are writing up some case studies. Will link as soon as they are available. Meanwhile anyone interested can access the AVS archive back to 2001 (http://www.actavetscand.com/archive) and can see that the numbers articles it has published per year has been steadily increasing.
July 7, 2012 at 9:26 pm
Thanks for that! Please drop in and comment again when the case-studies are available. (We see all comments on this blog, even on very old articles.)
July 7, 2012 at 9:40 pm
Hi Mike. Caroline Sutton and I maintain a list of society publishers of OA journals. (It’s a Google spreadsheet under a CC-BY license, open for public edits.) When we launched it in December 2011, it listed 530 societies publishing 616 full (non-hybrid) OA journals. Thanks to user contributions, it now lists 609 societies publishing 702 full OA journals. Here’s the spreadsheet [ http://bit.ly/oaj-society ], and here’s my article about it from December 2011 [ http://goo.gl/OMGpZ ].
July 7, 2012 at 11:16 pm
If the societies’ revenues from “old” (term to be defined) articles are not a significant part of their budgets, (I am not sure if that is correct) then I think it is a good idea for the societies to switch their journals to “delayed open access” (like, for example, Journal of Evolutionary Biology, which articles older than two years are open). At least as the first step.
…but probably I should just keep quiet because I have almost completely no qualifications to comment on the subject and I have no idea whether this idea make sense in the real world.
July 8, 2012 at 10:17 pm
^Tomasz…you hit it on the head…however, this might feel like the indexing wall JSTOR and PubMed users experience. Many articles aren’t indexed until a 3-5 years after their publish date. http://www.ebscohost.com/uploads/imported/thisTopic-dbTopic-226.pdf (sorry the link is Ebsco…lame, I know).
July 8, 2012 at 11:00 pm
Regarding a hybrid OA model for society journals, I do think it is at least the way to go in terms of a first step – especially if the society is worried about losing income from subscriptions or the academic publisher. But I do prefer green OA of non-typeset pre-prints in combination with delayed OA for final published articles as Tomasz suggests, and there’s no reason this shouldn’t be financially doable for most journals. I can only really speak about JVP in this regard, but:
1) I don’t think data about whether arXiv hurts subscription revenue is particularly relevant to whether converting a journal to gold OA will hurt subscription revenue (not that you were necessarily implying this). arXiv is a non-peer reviewed pre-print archiving service, and most journals I know of already allow you to post non-typeset pre-prints online after the article is published in the journal. It doesn’t appear that JVP’s copyright policy precludes this, though they also don’t explicitly allow it.
2) For a journal/society of JVP/SVP’s size, it is financially unfeasible to convert to gold OA if you want to maintain subsidizing free publication for society members. Of course, a society could launch a fundraising effort to build an endowment to do so, but that is a multi-year effort to raise serious cash.
3) If conversion to gold OA required paying a publication fee of $1K-$3K, I think there would be a major revolt by society membership. As people have pointed out, a lot of folks cannot afford this cost within the field of paleontology (though some people whinging about it are unaware that their institutions have funds available to offset OA publication costs), and society journals are small enough in volume that they can’t subsidize the ‘financially-disadvantaged’ scientists in the way that PLoS does with their waiver system. Ultimately, most people make practical decisions when it comes to financial resources; they cannot afford to be idealistic. So, if the rubber hits the road, I hypothesize that most SVP members would prefer free publication over paying for OA. If some folks have access to specific OA funding lines via grants or their academic institution, then that puts them on a separate track for decision-making about what journal to submit to.
4) With regards to the efficacy of a hybrid OA model, I think it can work well, but depends on two major factors: 1) how well you publicize the option; and 2) cost among peer journals. It seems to me that if your OA publication charge is $2K and below, that is seen as a decent value among journals with that option. Elsevier etc often have rather pricy OA charges, and furthermore, they don’t always publicize the option very well.
A great example of where a hybrid model works well is the Royal Society journals such as Proceedings B and Biology Letters. They do a good job publicizing the option, and you see a significant percentage of new articles (inc. paleo ones) where the authors have paid for OA. So it can work.
5) I’m skeptical about your statement “JVP‘s optional-OA uptake has been disappointing, too.” JVP has only offered an OA option for a few months. It takes time to publicize the option, and then have enough manuscripts pass through the system to get a statistically significant sample size. Furthermore, people pay this option after acceptance of their manuscript, so if people submitted their MS in the early spring, they may not get to the acceptance/OA payment stage until right about now, depending on the time it takes to get reviews back, number of times the MS needs revision, etc. So at least with JVP I think we need to be patient to evaluate success of the OA option.
Finally, what is the cutoff in terms of percent OA articles in a journal that you consider ‘disappointing’ vs ‘sufficient’ vs ‘encouraging’? Without presenting hard numbers, we have no way to evaluate your statement. There’s no cost to SVP to offer this OA option, so it isn’t a zero-sum game. Therefore, as long as more than 0% of authors take advantage of the OA option, I’d count it as a success. That said, I do think SVP should do more to publicize the OA option.
July 9, 2012 at 12:07 am
“no study has ever shown that Green OA…”
It’s a single example, but the top mathematics journal Annals of Mathematics (where, for instance, the proof of Fermat’s Last Theorem was published, taking up an entire issue) is published by Princeton University’s maths department, so not dissimilar to a society journal. They once tried becoming a OA journal, in that the electronic version was free but print was not. Subscriptions fell dramatically, even though it is a relatively inexpensive journal (it costs US$534 a year for combined print and electronic access). They have since reversed this position.
July 9, 2012 at 9:24 am
@David Roberts: there is a subtlety in the Annals of Math. example. They *are* Green OA, in the sense that authors are allowed to deposit their papers in arXiv. What failed was the attempt to become an *overlay journal* (the papers in their definitive form where *all* put on the arXiv). This seems to make little difference, but in fact mattered a lot. I think that overlay journals cannot expect to sell subscriptions, for green OA ones the matter is different (but this difference could vanish with the generalization of the use of arXiv; it has not yet).
July 9, 2012 at 9:39 am
Hi, Randy, thanks for a lot of interesting points.
I am not strongly opposed to hybrid in principle; but my growing sense that with a few exceptions it doesn’t work at all well. For example, see the results of Elsevier’s hybrid-journal program, “Sponsored Articles”, which in 2010 was taken up for 691 articles of the 250,000-odd that they published. That’s one in 360 articles.
Delayed OA always makes my skin crawl a bit. As an interim solution it’s not a disaster, but it always represents a compromise where something other than the progress of science is winning out.
It’s true that the example of arXiv is not a particularly close parallel to JVP. But it’s not easy to think of a closer one, since arXiv represents the systematic open-access availability of preprints for pretty much all the journals that it affects, rather than the much more sporadic pre-printing we’ve used to seeing in the biosciences. So I think arXiv may be the least bad analogy we have, much as elephants are the least bad extant analogue for sauropods.
It’s not obvious to me that that’s true. In 2010, JVP published about 156 articles (extrapolated across six issues from a count of 26 in volume 30 issue 6. In the same year, its publishing revenue according to its 990 form was $84,703. Even assuming all of that revenue was from JVP, which I guess is probably not the case, that still comes to only $543 per article. That suggests to me that JVP could continue on exactly the same financial footing if it were able to convert to all-Gold OA with an APC of $543 — about a quarter of the current discounted price of $2000.
And that of course would be on the assumption that all subscriptions were cancelled. If, as seems more likely to me, about half of subscriptions continued (because libraries want paper copies, or out of sheer traditionalism), then the Gold OA fee could also be halved to $271.
(Of course these calculations are extremely naive. I’ll welcome corrections from people who know more than I do about the finances.)
Agreed. But if the numbers above are right then the fee could be as low as $300 while still allowing some slop for 10% of submissions to take waivers. I guess that would not provoke much of a revolt, although with palaeontologists, really, who knows?
I take the more optimistic view that ideals factor into most people’s decisions, but that they’re not considered in isolation. If The Right Thing was to publish Gold OA but the fee was $1,000,000, then no-one would do it however idealistic they were. If the fee was $1 then anyone with a brain would do it, even if they were completely lacking ideals, for the citation advantage alone. (I’d argue that the same should apply at the $300 price-point, too.)
Somewhere between $300 and $1,000,000 lies each person’s switchover point — and it will be at a different level for different people depending on how much institutional funding they have, how idealistic they are, how much they care about the citation advantage, and so on.
It’s true, Proc. B does seem to be an outlier here, with a good, healthy elective-gold OA program.
I don’t think that can be right. It’s only been a few months since the price-drop to $2000, but the program has been available at least since the switch to Taylor and Francis three years ago.
Well, do we know of any JVP articles that have taken the elective-OA option in the three years it’s been available?
July 9, 2012 at 11:47 am
Mike,
Data for AVS, following its move to BioMed Central, shows (as you would probably expect) steady submission growth mirroring its increasing Impact Factor:
Year Submissions
2006 51
2007 65
2008 86
2009 125
2010 144
2011 156
2012 ~200 (projected)
Matt
July 9, 2012 at 11:49 am
Thanks, Matt. Actually, that’s not particularly what I expected: I thought perhaps the impact factor was growing because fewer articles were being published (lower denominator) while citations were staying fairly steady. Instead, we see that this is better news all round for the journal: the number of published articles is growing, but the citation rate is significantly outstripping that growth-rate.
July 10, 2012 at 12:31 am
Mike – a couple of comments.
Regarding my original point #2, I agree that if JVP charged a publication fee they could afford to make the journal gold OA. What I meant to say is that JVP could not afford this transition if they maintained free publication for members. Although there are more than ten times as many members as articles per JVP volume, membership money obviously goes towards many different costs. We also have alot of student members, and like most societies, professional membership fees effectively subsidize the cost of membership for students.
In terms of my point #5, it would really help if you and Matt actually talked to someone who knew about JVP before you posted stuff about it; alot of your errors would be rectified by fact checking with someone on the editorial board or publications committee. Your statement “It’s only been a few months since the price-drop to $2000, but the program has been available at least since the switch to Taylor and Francis three years ago” is somewhat incorrect.
Yes – T&F’s program has been available for three years, but it’s offered on a journal by journal basis; it is not offered automatically across T&F’s entire stable of journals. At the 2011 SVP Annual Meeting last November, I attended the Publications Committee meeting and confirmed that at that time, T&F did not offer an open access option for JVP; and I proposed that the journal should offer such an option. The committee agreed and T&F started the process of offering an OA option for the journal.
The first email to SVP members announcing the open access option was sent out on 23 January 2012, but it was buried at the bottom of the email about several different items:
In this email the stated OA publication charge is $2000. The first real publicizing of the OA option was in an email to members on 20 April 2012. This demonstrates a couple of things:
1) The JVP OA option has only been available for a little over six months, and it wasn’t really publicized until late April. Therefore, it wouldn’t be surprising if we’re still awaiting our first JVP OA article if people weren’t aware of the option until a couple of months ago.
2) There was never a “price-drop” for the JVP OA fee. It has always been $2000 for JVP since the OA option was first available in late January. True, this is a discount from what T&F charges for other journals, but T&F never started with a higher price for JVP and then dropped it to $2000.
July 10, 2012 at 7:44 am
Hi, Randy.
Not according to Andy Farke’s post of 15 October 2009 — a post that you commented on, saying “Certainly the cost for open access is steep – but don’t single out JVP on this. Quite a few non open-access journals charge a similar fee to make the article open”.
I don’t know — maybe we were all mistaken back then. It would be nice to refer to the FAQ that the society prepared at the time, which Andy mentioned in his post, but predictably enough it’s no longer there. FOr that matter, neither is the original press-release. *sigh*. Come on, SVP: Cool URIs don’t change.
July 10, 2012 at 7:50 am
I found a copy of the JVP’s T&F transition FAQ — they moved it to http://vertpaleo.org/PDFS/c1/c1f2e057-c8da-426e-aa94-bbe658e0a853.pdf.
It explicitly states:
July 10, 2012 at 2:45 pm
Just a few of points of clarification:
1. Randy’s account of our meeting with T & F is correct. The T & F rep at the meeting was apparently unaware that JVP could offer open access and had to confirm this with T & F before the OA option was granted. There was no existing advertisement of the previous arrangement on any official journal pages, only the statement buried in the more general document about the transition to T & F. Neither of the incoming Senior Editors were made aware of any OA option in existence when they joined the journal.
2. Since we’ve ‘explicitly’ been offering OA as an option there have been zero people who have taken advantage of it so far. In addition to mentioning the option on the journal webpages it is also a question asked during the submission process.
3. Following from the thread on vrtpaleo if you have any queries about the finances of the journal they should really be addressed to the Treasurer, Ted Vlamis. I’m not sure how much I can reveal of the various revenue streams in the public realm. Although the Society’s balance sheet is a matter of public record, details of individual commercial transactions might not be. However, such details are available to SVP members if they come to the Annual Business Meeting or enquire directly. T & F basically pay to produce the journal (production, marketing, distribution, etc.) and SVP receive a percentage royalty in return (based on institutional subscriptions, permissions, other sales, downloads from other platforms etc.).
July 10, 2012 at 3:54 pm
Thanks, Paul, that’s helpful.
It’s a bit disturbing that no-one seems to have even noticed that JVP had an open-access option! I suppose it shows that as recently as a couple of years ago, OA was a bit of a fringe issue. It’s great to see how much it’s moved into the mainstream in so short a time.
I’ve not written the JVP treasurer or other board members because I’m not actually a member myself. I’m working only with information that’s in the public domain. It would be great if someone who is a member could pick this up and look into it in more detail — I’m well aware that the coarse-grained information I’m working with and my own financial naivety make it impossible for me to draw any meaningful conclusions at this point. I’m trying to get the questions asked, rather than trying to propose answers!
Finally, of course I do understand that in general the specifics of the deal made with T&F may not be on the public record, and that you’re in a position to divulge them on a public forum. (I think they should be matters of public record, but that’s another issue.)
July 12, 2012 at 4:23 pm
For another take on this issue, see Open Access – What’s a Learned Society To Do? over on The Scholarly Kitchen.
July 12, 2012 at 5:28 pm
For comparison, you might be interested in the Geological Society of London figures (also a charitable, learned society) http://www.geolsoc.org.uk/webdav/site/GSL/shared/pdfs/Annual%20Reports/Click%20here%20to%20download%20the%20full%20accounts%20and%20financial%20statement%20for%202010.pdf
my assessment of the numbers (do correct me if my logic is mistaken, I’m no accountant):
(page 18) in 2010 publication income was £1,990,116
and publishing activity costs were £1,477,306
income breakdown (p39):
£660,000 from books
£1,330,000 from journals (other publications)
cost breakdown (p28):
£826,607 are ‘direct publishing costs’
and £466,569 are ‘staff costs’
and £254,977 are other publishing ‘Support costs’
so the GSL perhaps makes >£500,000 ‘profit’ per year if the publication income and costs can be validly considered independent of the other activities of the society (?). These funds are presumably spent elsewhere to support other non-publication related activities of the society.
I’ve used the exact numbers given lest I get accused of underplaying or exaggerating the numbers by rounding them. I just want to report what I believe to be factual ;)
It’d be interesting to get a tabulation of more figures from various different charitable-status learned societies that publish journals e.g the British Ecological Society … anyone want to start a spreadsheet or something?
July 13, 2012 at 6:00 pm
[…] is timely: two recent blog posts about OA and societies. One by Mike Taylor, one by the Scholarly Kitchen. There are other white papers etc also. I’ll hopefully get […]
July 16, 2012 at 7:12 am
[This is a comment I left at Tim Gowers’ blog, but would also be appropriate here.]
This might be of interest to some here: https://www.martineve.com/2012/07/10/starting-an-open-access-journal-a-step-by-step-guide-part-1/ Part one describing startup for a new open access journal. The up front and recurring open costs are detailed. There is also an estimate on time for copyediting, but this is in the humanities. I don’t know if this would be an upper or lower bound for mathematics, and the author has closed comments so I can’t ask about the assumptions behind this.
July 16, 2012 at 7:32 am
Thanks, David, this is useful.
July 17, 2012 at 12:03 am
Part 3: https://www.martineve.com/2012/07/12/starting-an-open-access-journal-a-step-by-step-guide-part-3/ with links back to parts 1 and 2.
July 22, 2012 at 3:58 pm
[…] What should society journals do about open access?(svpow.com) […]
November 16, 2012 at 8:46 pm
[…] doubt, there is an issue here — it’s the one potential downside of the shift to OA that bothers […]
March 28, 2013 at 1:26 am
[…] I would describe it “has been analysed in detail and the jury is in”. As noted previously here on SV-POW! and in my submission to the House of Commons, Swan’s data says […]
June 14, 2013 at 8:40 am
[…] not said much about the hybrid model here on SV-POW!. A year ago, when we were discussing what society journals should do about open access, I […]