The Matthew Effect derives its name from a Biblical parable in the Book of Matthew. The parable is best stated in modern language as, “the rich get richer, and the poor get poorer.”
Recently, the Wellcome Trust published files showing where and how much it paid for article processing charges (APCs) for open access (OA) publication.
The Matthew Effect seems to be present throughout the list.
In the year covered (2012-13, as their fiscal year is not a calendar year), Wellcome spent over US$6.5 million on OA publication fees covering a set of 2,127 articles, for an average of US$3,055 per article.
The minimum spent by Wellcome on an APC was US$75, while the maximum APC was nearly US$22,000, which was the APC for an OA book published by Macmillan, the parent of Nature Publishing Group. The highest article APC was US$10,000, which was for a publication called Public Service Review, a magazine apparently geared to policymakers in the UK. The publication may not be available any longer, as its web site is turning up missing. This was noted as well in Research Information, which states that, “it is difficult to find details of this journal and the URL listed for this journal in the Wellcome Trust’s document now appears to be available for sale.” Does Wellcome deserve a refund?
The next most-expensive APC (US$9,500) is for Lancet Neurology, an Elsevier journal. In fact, Elsevier dominates the top APCs, as does Nature Publishing Group. PLOS varies, with high APCs for its selective journals and low APCs for PLOS ONE.
Given all this, it’s not surprising that the publisher that benefited the most from this spending is not an OA start-up or fresh-faced publisher, but rather the largest scholarly publisher in the world — Elsevier — which pulled down US$1.65 million of the total, or 25%. The next biggest recipient is another large publisher, Wiley, which took in US$806,000, or 12%, followed by PLOS, which received US$581,000 (9%). Next was Oxford University Press, which netted US$514,000 (8%) — OUP’s APCs tended to be on the low side, it should be noted. These four publishers accounted for 54% of the OA APCs paid by the Wellcome Trust in 2012-13.
For PLOS, 56% of the APCs it received from Wellcome in 2012-13 went to PLOS ONE, at an average cost of US$1,550. The average revenue per article for PLOS overall was just US$1,880. Elsevier commanded a higher price per article, receiving just over US$4,000 per published article — a 30% premium on the population average, and well over twice what PLOS received on average per article.
These data confirm the belief that the OA business model will reinforce the consolidation that has already taken place in the publishing marketplace, as it rewards scale and size, which are highly correlated. It may also incentivize further consolidation. The acquisition of Frontiers by Nature Publishing Group is a point in favor of this perspective, as is the fact that BioMed Central is owned by Springer.
In fact, if you add Springer’s OA funding from Wellcome to BMC’s, you see that Springer is actually another large publisher receiving a high proportion of Wellcome funding for APCs — a total of US$528,000, putting Springer/BMC ahead of OUP in total.
Add it all together, we have five publishers netting 63% of the APCs expended by Wellcome Trust in one year. Here is the resulting list of the Top 5 recipients:
- Elsevier — US$1.65 million, or 25% of Wellcome’s OA APC spending
- Wiley — US$806,000 (12%)
- PLOS — US$581,000 (9%)
- Springer/BMC — US$528,000 (8%)
- OUP — US$514,000 (8%)
Nature Publishing Group wasn’t far behind, taking US$330,000 in APCs from Wellcome Trust during 2012-13. It’s average APC was US$4,600.
When it comes to journals, PLOS ONE was the big winner, receiving almost as much as Nature Publishing Group in APCs from Wellcome Trust during the same period (US$327,000). The next biggest journal recipient was the Journal of Biological Chemistry, with US$122,000 in APCs from Wellcome. These journals represent major outliers in a study of citable objects done last year in a post by Phil Davis, suggesting once again that OA performs well as a quantity-based business model.
No naming authority was imposed on the file, so entities have a variety of names. For instance, the American Chemical Society is referred to in a number of ways — ACS, ACS Publications, American Chemical Society, and The American Chemical Society, among others. PLOS ONE is listed in a similar variety of ways, as are most of the publishers and journals with multiple entries. A number of articles are listed as the composition vendor, making analysis even more complex. By allowing authors to freestyle the name of the payee, we are creating a very loose data source for analysis.
It’s interesting to contemplate exactly what Wellcome bought with its US$6.5 million, as many subscription journals in the fields covered are subject to 12-month embargoes. One could argue that Wellcome paid an average of US$254 per month per article to make the articles free early. Put even more starkly, Wellcome is now paying the equivalent of US$542,000 per month in aggregate this year to make these 2,127 articles free for the 12 months we’re in, rather than paying no APCs and allowing the articles to be published in subscription journals that honor 12-month embargoes.
One prominent OA journal missing from the list of APC recipients is PeerJ. I don’t know what it means, but active speculation about PeerJ’s future is not encouraging, and this absence from Wellcome Trust’s list of APC payments may add to existing concerns.
In 2012, I speculated that OA had the potential to exacerbate the Matthew Effect, writing:
. . . given the strong tendency of scholarly achievement to drive the Matthew Effect in funding, in publication, and in impact, it seems that OA publishing’s future will be as likely as not to perpetuate the problem and, if not thought about carefully and implemented judiciously, may actually exacerbate the “rich get richer” predilection of academic institutions, individual researchers, and prominent hypotheses by coalescing around current sources of funding and prestige, and inflating them through its business model.
Judging from these data, it appears that APCs are increasing and finding points of differentiation, with established publishers commanding much higher prices than newer, pure OA publishers — suggesting yet again that the Matthew Effect is absolutely a part of OA’s future.
37 Thoughts on "Wellcome Money — In This Example of Open Access Funding, the Matthew Effect Dominates"
“A number of articles are listed as the composition vendor …”
What is a composition vendor?
Just to let you know – the publisher of “Fungal Disease in Britain and the United States 1850-2000” is Palgrave Macmillan, the humanities and social science publisher, and sister company of Nature Publishing Group.
This was the first open access monograph to be funded by the Wellcome Trust, and the APC was £11,000 plus VAT.
Corporate Communications Manager
You state, “One prominent OA journal missing from the list of APC recipients is PeerJ.”
Note that for the academic year 2012-2013, which this Wellcome data covers, PeerJ had only published a handful of articles out of the ~400,000 Open Access published in that period. Approximately .01% of all Open Access articles may have been PeerJ as it had just established operations (publishing started 12th Feb, 2013 and most of the year was free of any article charges). When did the 2012-2013 academic period end in 2013?
So, statistically it’s no surprise that paid PeerJ articles didn’t show up on this list of just 2,400 articles (.01% of 2,400 is 0.24 articles expected at most – again many were free of APC). For that matter, eLife, F1000Research, and any other journal also started in 2013 or late 2012 would likely not appear on the list.
What is surprising, perhaps, is that despite just starting operations, PeerJ seems to be on the minds of quite a few people. Case in point, that you thought to include PeerJ in your analysis. Certainly we hope to earn their trust and expected an understandable wait and see attitude from skeptics in the early days.
Couldn’t one put together similar order of magnitude numbers for most of the journals on this list, PLOS ONE aside? What percentage of OA articles did “Acta Ophthalmologica” publish in the financial year listed for example?
David – I can’t speak for other journals. However, for PeerJ there were just a handful of paid Open Access articles that were published between Feb 12th, 2013 and before the end of the UK financial year (April 5, 2013). I believe < 10 were paid, all the rest free in that period. Additionally, a search  on PeerJ shows that Wellcome Trust backed researchers are publishing there, but there is a mix of funds being used (when the article wasn't free to publish), which again would need to be taken into account. So, again statistically we wouldn't see PeerJ articles appearing until next year's Wellcome Trust list, and even then it would be underreported.
Good point–like most OA launch journals, PeerJ has done several periods where no APC/Membership fee is charged. Anything published in these periods wouldn’t show up on a financial account like this.
But the point I was trying to make (poorly articulated) is that perhaps given the small sample size here, one could say it’s surprising to see any particular journal (PLOS ONE aside) included.
The breakdown didn’t startle me so much, I admit…
Of a sample of 305 varied earth-sciences/environmental-sciences/biology papers I have to hand, mixed between gold and non-Gold, covering my organisation’s recent output with no filtering – 83 are Wiley (inc. AGU & AMS), 60 Elsevier (inc. Cell) and 28 Springer, or 27%, 20% and 9%.
Now, I work with researchers with a great fondness for publishing in a particular large journal, and this is what skews it so strongly to Wiley, but even so – that’s a greater concentration from just two publishers than Wellcome are seeing from three. (Springer is actually #4 with us – the third-place spot is Copernicus/EGU, 30 papers, rather than PLoS, 17.).
The Wellcome data is also looking at a more restrictive set than the sector as a whole because (of course) they only pay for gold OA in journals which offer gold OA, which may give an apparent advantage to the large publishers who moved quickly and set up blanket policies.
Finally, PeerJ – as I understand it – works on a “membership” model where an individual researcher pays for membership, getting an entitlement to publish a set number of papers. It might be that this does not qualify to be reported as an APC or is more likely to be reported under another category (I’m not familiar with the Wellcome’s system) but equally it might be that any researcher publishing in PeerJ has *already* paid for it.
Incidentally, there’s a (somewhat) tidied dataset at https://docs.google.com/spreadsheets/d/1RXMhqzOZDqygWzyE4HXi9DnJnxjdp0NOhlHcB5SrSZo/edit#gid=0
The highest article APC (for Public Service Review) is a publication which is associated with some controversial sales tactics (http://www.andrewjaffe.net/blog/2010/09/public-service.html) and whose parent company appears to have gone bust last summer (http://www.stokesentinel.co.uk/200-jobs-axed-Newcastle-based-publisher-collapses/story-19704837-detail/story.html). I am not sure that it counts as a peer-reviewed journal…
So what lesson do you draw from this? If APCs aren’t well-monitored, funded researchers could be spending significant amounts of Wellcome or other funders’ money on APCs with fly-by-night publishers? Why was this allowed? This one fascinated me for exactly the reasons you state — the publisher got a lot of money, seems to have vanished, and wasn’t offering publication in a peer-reviewed journal. Yet, they were paid.
Something I was surprised by was the very high number of hybrid journals listed. Where there is a listing for Pure OA or Hybrid, the latter make up 43% of the journals used, and by eyeballing the unlabelled ones, there’s a significant fraction of hybrid journals to be found there as well.
Hybrid OA journals have been written off by many as a failure, but here we’re seeing them play what seems to be an important role for Wellcome authors.
Hybrids seem like a natural compromise and a way for the market to sort itself out. Mind you we are talking about a movement that will take a decade to work through so this little snapshot should not be taken too seriously. It is no surprise that the popular publishers and journals remain that way for now, that there is no APC price war, etc. When or if major shifts occur it will be slowly. So diagnosing a Matthew effect from this snapshot may be premature.
With Wellcome building OA costs into grants, why would authors be sensitive to price? I suspect Impact Factor (or IF’s more general cousin, “prestige”) still rules their decisions.
What is going to happen when some bean counter at Wellcome (or a government agency following a similar policy) decides to enforce some economy? Bidding for papers?
There is no reason why the APCs cannot be contained to the minimum of the charge in this Internet World provided the Researchers become more open minded and not Achievement Centered. For reporting of a finding is itself important enough than how it is done. Imagine the day and the group of thinkers who had arrived at the IDEA of A WHEEL and then imagine of the 21st. Century without the WHEEL. Can anyone please name the Journal and the Authors who published their finding first ?
WELLCOME TRUST will, therefore, be better advised to spend that US$6.5 M instead on Research Projects more demanding to the uplift of the down trodden/Health Research than confirming to the Matthew’s Effect with such a heavy uncalled for payment instead spend only about 1% of that expenses now on by publishing in the http://www.ijsmet.com (International Journal of Science Medicine Engineering & Technology) which, of course, will speak a lot more louder about the spirit of Mr. Wellcome,s purpose of creating the Trust. Also, do please note what the Editor-in-Chief of eLife A Nobel advocates for now. Let us, therefore, take a NEW LOOK in THIS NEW WORLD of Global-Village-Age.
As far as I’m aware, the Wellcome Trust neither encourages publication in “vanity” journals, nor performs evaluations based on broken publishing venue criteria.
However, the researchers that it funds are being evaluated – by their current and potential future employers – using such broken measures as the journals they have published in.
Fix the broken evaluation mechanisms, and the funders (and researchers) will most likely adapt.
Open access isn’t about which companies benefit and which ones don’t. It’s about content being available for everyone to read and reuse and, as you make a great job of showing, it’s about visible costs.
The fact that you are able to try to shock your people about the Wellcome spend is a good
thing. Difficult to do under the legacy model, when costs to the research budget are largely
Open access isn’t about which companies benefit and which ones don’t.
I suspect that this depends very much on who is talking about OA and about how up-front they are about their agendas. Some players see OA as a quick money-maker for themselves; others see it as a way to pull market share away from toll-access publishers. Neither of these groups is likely to say that out loud, but it seems clear to me (for what that’s worth) that such agendas are nevertheless a significant driver.
Kent, I think you have identified the wrong rogue publisher to highlight. What I find far more concerning than $10k possibly misappropriated by a fly-by-night publisher is that Elsevier has accepted $1.65 million in payment but not delivered on their side of the deal by actually publishing the works in an Open Access format.
As people like Mike Taylor and Peter Murray-Rust (eg ) have highlighted repeatedly – for at least two years now – many of the works Elsevier have been paid to publish OA remain behind paywalls, and the majority of articles on the Wellcome list published by Elsevier are doen so with the statement “Copyright Elsevier. All rights reserved” – no mention the work is Open Access, no mention of which CC licence it is published under, nothing.They have had years to sort this out and, as these figures show, have received a lot of funding through this route – but still they haven’t bothered. Should the Wellcome Trust (and other funders) now sue them for breach of contract do you think?
The link to Peter Murray-Rust’s blog, where many of these incidence are discussed, seems to have been lost from my comment above – it is: https://blogs.ch.cam.ac.uk/pmr/
For what it’s worth, Elsevier has responded here:
A good insight into the systems problems inherent in providing hybrid OA. It appears that Elsevier has not charged anyone for APC content but CCC et al may have mistakenly collected $4k or so. If so then claims that Elsevier is selling APC content are simply false.
If an article is meant to be OA but is actually behind a paywall there are at least two types of victims:
a. the people who incorrectly paid for the content
b. the people who didn’t access/reuse the content because it was behind a paywall.
It is the later group that the research bodies are paying apcs to supply. It is the former group that that Elsevier has identified, and are looking to recompense, in their response.
If the Wellcome Trust or any other funder is paying thousands to pounds to have an article freely available and re-usable then, you know what, it should be – and the conditions for it’s use and reuse should be transparently and explicitly stated. If that doesn’t happen – it IS misselling. QED.
If I pay for a few thousand pounds to upgrade and have, say, a 4×4 drive option included on my new car, and I then find that it doesn’t have that option, I would have every right to be both angry and compensated. An argument that it is difficult to install that option (or even that some dealers find it difficult to install that option) wouldn’t go terribly far as a defence – and nor should it.
Excuse me Rupert, but according to Elsevier no OA articles were behind a pay wall. A few people did mistakenly pay royalties for reuse of OA articles because the CCC button was there. If you have evidence to the contrary let’s see it. I have no dog in this fight.
Recall Elsevier offers authors the choice of CC licence applied to the work, and in many cases charges different rates for different licences.
So – some evidence:
On March 21 2012 – almost precisely 2 years ago – Mike Taylor pointed out the problem in a blog post – with screenshots etc all the way through.
On August 12 2013 – seven months ago – Peter Murray Rust pointed out problems of Open Access articles sitting behind paywalls
On Dec 18 2013 Stephen Curry (@Stephen_Curry) tweeted:
“.@ElsevierConnect – we paid you >£2k for #OpenAccess http://www.sciencedirect.com/science/article/pii/S0166093412003916# … But PDF says “© 2012 Elsevier B.V. All rights reserved”.
On March 11 2014 Mike Taylor summarised the situation in another blog post
Last week I looked at a sample of Elsevier published articles from the list published by Wellcome Trust – every single one was marked “Copyright Elsevier. All rights reserved”. Every webpage had a link to “permissions” which took you to a CCC site to receive a quotation for reuse of the work – but there was no indication anywhere that the work was open access, or which Creative Commons licence had been selected by the author and paid for by Wellcome. As a test I requested permission to reuse an article (“The Emergence and Representation of Knowledge about Social and Nonsocial Hierarchies” in the journal Neuron) and received the quotation of USD $927.36.
Today I have checked some more. One thing has changed – the permissions link has gone. So – for example – the first article I selected today was:
“Epigenetic factors influencing resistance to nuclear reprogramming” published in Trends in Genetics – for which an apc of (GBP) £2368.11 was paid.
Place that title into Google and you get links to the Journal page and PubMed central
The Journal page is:
with a clear statement “Copyright © 2011 Elsevier Ltd All rights reserved.” Absolutely no mention that it is an Open Access article – or which CC licence it is published under, and no way for me to discover this information.
The PubMed central page is:
Once again we find the same “Copyright © 2011 Elsevier Ltd All rights reserved” statement, no mention of the CC licence, and absolutely no mechanism for me to discover which licence.
I have not gone through every Elsevier published article on the Wellcome list – but what I describe above is repeated for the vast majority of the articles I have looked today as well.
But there are some that do slightly better, for example:
“Assessing the mechanism of response in the retrosplenial cortex of good and poor navigators” published in the journal Cortex
is clearly marked as Open Access – however the copyright information at the very end of the article reads “Copyright © 2014 The Authors. Published by Elsevier Ltd. All rights reserved.” and again there is no obvious indication of what CC licence it is published under, or how to find this out.
(If you persevere and go to the pdf edition you can find the following additional information “This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/).”
So here is a conjecture that may just be true: Despite having being informed of these issues over 2 years ago, Elsevier has failed to fulfill the publishing contract it signed with authors in every single Open Access article on the Wellcome list.
What you are describing is exactly the copyright notice problem that Elsevier says it is trying to fix. But in no case that I can see is any OA article behind a pay wall, which is what you originally claimed. Do you have any evidence to the contrary or have you just changed your claim?
That hybrid journals should create copyright confusions is not surprising. Elsevier says they have been working to fix the problem and I believe them. This is like a recall in the car world. It is not fast and it is not pretty but it is honest. Elsevier is not a person, it is a very big company, so “being informed” takes quite a while.
No David, I have not changed my original claim in any way – please do re-read what I said in those first two messages. There have been throughout two points of concern a. that many articles that should have been free to read have been published behind a paywall (as the blog posts by Peter Murray Rust demonstrate) and b. that the majority of the articles on the Wellcome list are published with an ‘all rights reserved’ statement (I hope that is well documented above)
And No, this is not like a product recall when a fault is discovered. The ‘product’ being sold by Elsevier to academic authors is an OA publication under an author selected CC licence. For over two years Elsevier has been aware that there was both a problem with their licence statement and no way for a reader to ascertain the CC licence or re-use rights associated with a paper. Subsequently (the Wellcome figures are for the 2012-13 year) they sold millions of dollars of this ‘product’ to thousands of academic authors (and not just Wellcome funded authors) in the knowledge that there was a problem and that they were not actually able to supply the ‘product’ being purchased. That is the behaviour of, at best, a rogue – which is where I started.
Better metadata will help track funders and OA. An example of what CrossRef is starting to collect is here – http://api.labs.crossref.org/10.1093/nar/gkr785.xml – sorry for the XML but it shows we have funder names and licenses. In addition, there is a pointer to the full text for text and data mining (“tdm”). What CrossRef doesn’t collect is any information on APCs that have been paid. Here is another example – http://api.labs.crossref.org/10.1103/physrevb.84.214434.xml – where there are different licenses for different versions and information about archiving arrangements since DOE funded the research that the article reports on. This type of metadata will drive the CHORUS services.
I agree that this type of metadata is absolutely essential to make CHORUS viable. The question is will there ever be enough to do that? My understanding is that FundRef is being supported by very few publishers. If federal funding agencies have to get most of their articles from the authors then FundRef and CHORUS do not help them. In fact it becomes an extra burden to sort out which funded articles are and are not in FundRef/CHORUS.
For what it’s worth, Cambridge released their data on APC payments made for OA articles for compliance with the RCUK policy. On first glance, their numbers parallel those of the Wellcome Trust, with about 25% of total payments going to Elsevier: