Show me the data!

Seeking Justice for Readers

February 27th, 2017 | Posted by rmounce in Paywall Watch | Wrongly selling OA articles - (Comments Off on Seeking Justice for Readers)

I am highly curious as to why Elsevier do not seem to be responding to emails at the moment:

Four days ago, continuing an existing thread on the public GOAL mailing list, I wrote to Dr Alicia Wise (Director of Access and Policy at Elsevier), about how Elsevier’s paywall systems are wrongly defrauding readers across the world by charging them to access content that has been paid-for to be open access.

Below is the full content of my message to the list, nicely formatted for clarity. I re-publish it here because I am still seeking answers to my questions, and I still want justice for the many unknown readers who may have been defrauded by paying to access content that should instead have been “open access” to all.

I remind readers that the scale of this is not trivial at all. Elsevier themselves admit in 2014 they attempted to refund or credit “about $70,000” to readers who had been wrongly charged for open access content.

Dear Alicia,
Approximately five days ago you wrote on behalf of Elsevier to this mailing list in response to my finding of a single paid-for hybrid “open access” article paywalled and actively being sold at the Elsevier journal Mitochondrion.

In the response Elsevier sought to re-assure the world (open access is for the benefit of everyone) that:

“We’ve gone through the system, this is the only article affected.”

I then found another paid-for hybrid “open access” article paywalled and actively being sold at the Elsevier journal The Lancet.

We appear to have had no official response from Elsevier since.

Today, an independent analysis by Christoph Broschinski of more paid-for hybridOA articles at Elsevier journals may have found up to five additional paywalled, for sale articles. The Cambridge one is a mistake, the payment was for page charges & colour figures not open access, I have done due diligence on this and checked myself since I am at Cambridge.

I’m struggling to reconcile what I have found and what Christoph might have found with Elsevier’s statement:

“We’ve gone through the system, this is the only article affected.”

The most likely conclusion I can draw from Elsevier’s statement and these reports that appear to conflict with that statement is that Elsevier’s system is not adequately tracking paid-for hybridOA articles. 

Assuming this to be true:

1.) Will Elsevier openly publish on a single web page, on a continuous, ongoing basis, the exact DOIs of all articles that Elsevier has been paid to make “hybridOA” , including the DOIs of articles that Elsevier were paid to make open access, that now reside at journals published by other publishers (if the journal was subsequently transferred to another publisher) ?

This will enable any interested party to:

a) Check that each and every one is actually freely accessible from the publisher site landing page

b) This ‘master list’ of Elsevier hybridOA can be cross-checked against institutionally-held lists of paid invoices. Any articles listed by an institution as paid-for OA, but not on Elsevier’s hybridOA ‘master list’ can be further investigated, to perhaps further reveal more articles that should be “open access” that Elsevier’s faulty “system” has overlooked.

2.) Will Elsevier refund 100% of the paid APC to each institution, funder, or individual that has a wrongly paywalled paid-for “open access” article behind a paywall?

3.) Will Elsevier hire and fully pay for an independent 3rd party forensic accounting firm to go through their pay-per-view and re-use licensing data/systems and records, including the period from January 1st 2005 until today (23rd February 2017), to produce a thorough openly available report on the extent of PPV payments AND re-use licensing payments for articles that should not have been sold to access, or to re-use?

I hope Elsevier will do this to ensure that every individual who has paid to access or re-use ‘mistakenly’ paywalled “open access” material is refunded in full, with interest, including the local taxes applied not just the article fee, and not with “credit” that can only be used to purchase other Elsevier goods or services.

4.) What meaningful assurances can Elsevier give that it will not make these mistakes again, given that it appears to be making these mistakes over and over again?

Full open access publishers have an error rate of precisely 0 out of over >500,000 articles published so far e.g. PLOS, PeerJ and eLife.

Errors such as these are simply intolerable and have the potential to cause great harm.
For instance, imagine if an article providing the first report of Ebola in Peru was paid-for to be hybridOA but was instead mistakenly paywalled…
Since Peru is now “too rich” [1] to qualify for HINARI, but still too poor to pay to subscribe to most subscription journals, many Peruvians would not have access to this vital information unless it was open access.

c.f. the first report of Ebola in Liberia which was also infamously paywalled at an Elsevier journal [2]




Remarkable ongoing chaos at OUP

February 21st, 2017 | Posted by rmounce in Paywall Watch | Wrongly selling OA articles - (Comments Off on Remarkable ongoing chaos at OUP)

Just a quick post tonight. Yet another paywalled open access article. It’s becoming a daily occurrence!

The total chaos that has been ongoing at Oxford University Press journals over the pass month is still ongoing would you believe it?

Having thoroughly shamed Elsevier for paywalling and selling paid-for open access articles it would only be fair to pour scorn on OUP for doing the same thing too at the journal Systematic Biology. The article ‘Bayesian Total-Evidence Dating Reveals the Recent Crown Radiation of Penguins‘ first published on 1st August 2016, has a landing page which clearly states it is “an Open Access article”, yet below this it also says “You do not currently have access to this article” and further below that you can pay $42 to be granted 24 hours access to this article. Like I said before, hybridOA at many publishers is seriously broken!


The TL;DR summary: In February 2017, when Elsevier were accused of selling one paid-for hybrid open access article, at first they sowed doubt about it, then three days later admitted it to be true. In their admission they state that it is the only wrongly paywalled open access article “affected” at their websites. They have apparently checked their “system” and say there are no more wrongly paywalled articles at Elsevier websites. Their statement is however demonstrably incorrect as I show in this post. There are more paywalled open access articles at Elsevier, and by their own admission Elsevier don’t seem to know about them. This is professional incompetence. HybridOA as whole, not just at Elsevier, is not reliable. We should not use hybridOA services if we actually want reliable open access in perpetuity. 

Some Background

Elsevier and other legacy publishers, have an unfortunate track record of paywalling articles that research funders have paid-for to be openly accessible. Elsevier in particular are notable for their consistency in this regard: they have been caught by diligent readers, selling access to “open access” articles in 2014, 2015, 2016 and 2017. They are so notable for this practice that it is on their Wikipedia page amongst a long section of other controversies.

Responding to the March 2014 scandal, Elsevier claimed it was merely “bumps on the road to open access”. Shortly before Christmas (a great time to bury bad news) on the 22nd December 2014 Elsevier quietly disclosed that they had refunded “about $70,000” to readers across the world who had purchased access to articles that should have been free to access at Elsevier. Naturally, as this self-disclosure came just before Christmas it went largely unnoticed and unreported.    

Now we fast-forward to February 2017. As an academic, fed-up with abysmally poor publishing services, I have decided to systematically run heuristic checks on articles that Wellcome Trust has paid-for to be open access. The Wellcome Trust and other research funders have started to publish detailed reports on which articles they have paid-for to be open access, with which publishers, and how much they paid publishers to make each article open access in perpetuity. I have used these public records supplied by Wellcome Trust to cross-check articles that are supposed to be open access against what is actually open access at publisher websites.

From this exercise so far I have found at least seven Wellcome Trust paid-for “open access” articles that are currently paywalled, and for sale to readers for up to $58.80 per article, at major academic publisher websites. I have publicly given details of one of these articles so far: paywalled at Elsevier at a journal called Mitochondrion.

The first response: sow doubt on the claim

Initially, Elsevier responded to my observation on a public mailing list:

Ross – we are looking into this, but an initial glance at the article suggests that the author first decided to publish gold OA and then decided against it.  We’ll have a little look to make sure all the metadata is correct.

— Dr Alicia Wise, Director of Access and Policy, Elsevier | Feb 15th 2017 09:21:44 GMT (source)

Note that Elsevier’s initial response suggests that I might have been incorrect in my assertion that this article should be open access and should NOT be on sale to readers.

I am a scientist. I make statements based upon evidence. I am no fool. But Elsevier seem to think I am.

The second response: admit to one error but clearly deny any further extent

Three days later Elsevier admitted that I was correct – the Wellcome Trust had in fact already paid Elsevier £2,168.08 many years ago to make this article open access, it should have been open access at their website, and should NOT be on sale to anyone.

Hi Ross, this has taken a little time to bottom out but indeed the article should be OA. In 2013 you may recall we initiated an exercise to add/correct and generally clean up OA license information on articles.  This article was part of that exercise, and a little unusual: we had invoiced the author, and then the institution wanted to receive the invoice but in a different currency, so the original  article invoice was unpaid during the cleaning exercise.

Our normal practice is to publish articles OA, but to move them back behind the firewall if after a grace period the invoice remains unpaid.  So it was picked up as an article with an unpaid APC and moved behind the firewall.  We’ve gone through the system, this is the only article affected. There’s nothing malicious about what happened here, and we’ll be reimbursing the institution’s APC, and your PPV charge.  We will also run a check to see if anyone else paid PPV for this article and if so reimburse them as well.

— Dr Alicia Wise, Director of Access and Policy, Elsevier | Feb 18 2017 00:03:24 GMT (source)

Within the above statement from Elsevier there is something much more astonishing. Elsevier clearly wants us all to “move along, nothing to see here” by stating that the Mitochondrion article is “the only article affected” by paywalls that shouldn’t be.

But there are more paywalled open access articles at Elsevier journals

Until now I haven’t disclosed publicly that there is at least one other paid-for “open access” Wellcome Trust article that is currently paywalled at an Elsevier journal, despite having been handsomely paid-for to be made open access in perpetuity. It’s at a rather well-known Elsevier journal too; The Lancet. According to data Wellcome Trust has made available, in 2012 they paid Elsevier £5,280.00 to make the article in the screenshot below, open access in perpetuity. In 2017 I find it is on sale at ScienceDirect for $35.95 + VAT. I bought access to it just to verify that it is actually on sale and indeed it is.

The screenshot below is my receipt documenting my purchase of the above “open access” article from The Lancet:

Now let’s return to Elsevier’s most recent statement and you’ll see why I’m shocked:

“We’ve gone through the system, this [the Mitochondrion article] is the only article affected.”

I have just provided evidence (The Lancet article) that there are more paid-for “open access” articles at Elsevier that are paywalled and have been sold to readers. As my receipt demonstrates, my purchase of this Wellcome Trust paid-for “open access” article post-dates the above statement by Elsevier. Therefore Elsevier’s statement would appear to be incorrect.

So we now have evidence of two further important things to note:

  1. Elsevier’s entire system for handling hybrid open access is broken
  2. Elsevier are evidently incapable of accurate self-assessment

Elsevier say they have gone through their “system” and that there are no other open access articles “affected” by paywalls. But I have just provided clear evidence to the contrary.

To be clear I’m not accusing Elsevier of knowingly lying about this. I am going to charitably interpret this as an unknowingly false statement – they believe it to be true from what their faulty “system” tells them, but in reality it is not true, other articles are affected. It is abundantly evident to me that Elsevier’s internal systems and processes for tracking invoices for APC payments, handling, and marking hybrid open access articles simply aren’t robust. I believe that this forms part of a body of mounting evidence that Elsevier is professionally incompetent at hybrid open access publishing and that this incompetence financially benefits them through pay-per-view payments & re-use licensing to the detriment of the scholarly poor (those outside paywalls). I don’t allege malicious intent. The point is that regardless of intent, this incompetency is demonstrably generating profits for them – that is undeniable and that it is causing real harm (financial loss to readers who have paid for articles that should not be for sale).

What is the true extent of this problem at Elsevier in 2017? Why does it matter?

I have to admit I don’t have a definite answer as to the extent, millions of articles are published every year, I can’t check them all myself. But one thing I do know and can now evidence, is that clearly Elsevier don’t know either! I think a full independent audit with forensic accounting is urgently needed. Why? As I mentioned earlier, back in 2014 Elsevier self-disclosed that they had charged readers $70,000 to access articles that should have been “open access”. Since 2014 the rate of publication of hybrid open access articles has increased so I would expect if a competent independent audit were done of purchases of ‘mistakenly’ paywalled articles between 2015 and now, the sum to be reimbursed might well be in excess of $100,000. Given Elsevier doesn’t know that some of its paid-for hybrid open access articles are on sale, can we trust their self-reported audit of such incidences prior to 2014? I think not, hence a thorough independent audit is needed.

The scale of this could be very big. I have only partially cross-checked the accessibility of hybrid open access articles from just one research funder at Elsevier. According to CrossRef data there are at least 14,155 different research funders. The charge to make articles hybrid open access at Elsevier is typically around $3000 but can be much more as we have seen from The Lancet article which was £5,280 to make open access. When found to be in error, Elsevier typically refund the APC-paid, so therefore if just one “open access” article per research funder is mistakenly paywalled, it would equate to a refund of approximately $42,465,000 USD. This amount is still chump change to Elsevier: a company that in 2013 posted an annual adjusted operating profit of £826 million on an annual turnover of £2.126 billion.

Nor in terms of scale is this scandal confined to just Elsevier. There are paywalled “open access” articles for sale at Emerald Publishing Group (exonerated 2017-02-27), Cambridge University Press and Wolters Kluwer / LWW to name but a few. I have evidence, which I have already informed Wellcome Trust of. I don’t make these claims lightly.

I stress however that this scandal is confined to hybrid open access publishing.

Only fully open access publishers are reliable

How many times have PLOS accidentally paywalled an article? [0]

How many times have PeerJ accidentally paywalled an article? [0]

Full open access publishers such as PeerJ, PLOS, MDPI, Hindawi, Copernicus, Pensoft, F1000Research, and Ubiquity Press to name but a few, have an open access compliance “error” rate of absolutely ZERO, out of hundreds of thousands of articles published so far.

In some respects these are facile statements to make about full open access publishers. But that is my point. It is known that PLOS and PeerJ cannot paywall and ‘mistakenly sell’ articles to readers because their systems are open by default and do not have paywalls. At fully open access publishers, all articles are open and thus when an APC is paid or a fee waiver awarded, there is never any failure to make articles available in a manner that is both freely accessible to readers (not paywalled) and licensed under an open access definition compatible Creative Commons licence that Wellcome Trust requires.

If research funders and institutions really want open access publication, under open access definition compliant licensing, they should use full open access publishers.

Luckily not all research funders allow their money to be used to pay for hybrid open access charges. This is sensible policy in my opinion, given the evidently incompetent open access service provided by many, not just Elsevier.

Conclusions: what can funders & APC-paying institutions do about this going forwards?

As soon as reasonably possible, stop paying for hybrid open access. It is not reliable. If funders and institutions really want open access, use preprint servers and full open access journals. Hybrid open access is demonstrably unreliable, vastly over-priced, and does not help towards long term goals of open access as it helps keep subscription-access journals going and takes good articles away from full open access journals. It is time to end the experiment.

TL;DR summary: ESA data papers should be free to read but Wiley (ESA’s new publishing ‘partner’) just charged me $45.60 yesterday to access one of them. They have done this kind of ‘accidental’ profit-generation before, as have other big publishers.

John Wiley & Sons (whom I will refer to as ‘Wiley’ from now on) is not a very competent company when it comes to providing free or open access to research. Don’t take my word for that. Ask the Wellcome Trust: over 50% of articles that they had paid to be open access with Wiley were not compliant with their open access policy. I have also had my problems with Wiley: I caught them selling access to thousands of articles that should have been free to access this time last year. They also paywalled an article I wrote which should have been free to access.

Despite all this, and the detailed letter I sent to the Ecological Society of America (ESA) back in 2013 during their open access consultation process, the ESA decided to switch to publishing with Wiley: a profit-driven company who’s goals conflict with the goals of the society. I was very disappointed with this decision.


Now that the switch is complete there are some problems readily apparent. Wiley are selling some bits of ESA journal content for $45.60 (inc. tax) a time that ESA did not previously charge readers to access. I discovered this yesterday on Twitter thanks to Jaime Ashander & Stephanie Peacock. So I made a test purchase to see if Wiley really were charging for access to this free content (they were!). Below are tweets documenting this:


Amusingly, the first time I tried to buy access to the article, my bank blocked the transaction thinking it was a suspicious payment to a scammy company! Only after I confirmed with my bank was I actually allowed to purchase access to the data paper – it really IS hard to access research that is paywalled, even when you have the money to pay for it!

ESA have acknowledged the problem on Twitter and will see if I can get a refund on Monday:


There is more than meets the eye to this case.

Data papers are still a fairly new concept to most. Thus I honestly didn’t know what I’d be getting from behind the paywall when I paid for access – I did expect more than just the abstract. It would not surprise me if others could also make this mistaken assumption (we are wearily used to abstracts hiding much longer papers behind paywalls).

Charging the authors of ESA data papers $250 with the excuse that this is for “long-term hosting and maintenance” is absurd and unjustifiable. At the very most it should be $120 which is what Dryad charges, with a reminder that Figshare and Zenodo continue to sustainably archive data for free. Charging each and every reader outside the paywall in addition to this $45 to read the abstract of an ESA data paper in PDF format is just ridiculous.

The cost of single-article purchases has now more than DOUBLED since ESA moved to Wiley. Below is a screencap I took from the old ESA publishing platform. ESA articles were paywalled for just $20 and that allowed 30-day access. Now with Wiley, the exact same content is available to me for $45.60 (inc. UK tax) and I only have a 24-hour permitted-access period. This price-hike and narrow access window are utterly absurd and unjustified. Is it any wonder everyone uses SciHub these days?

Does this help raise the awareness of ecological science?

The old paywall was half the price and gave 30-days access, not just 24 hours!






I’m also frightened that ESA had no idea this was going-on. This is exactly what happens when you give all your content to an unscrupulous oligopoly publisher like Wiley to sell on your behalf. It seems to me that many academic societies are receiving big fat cheques every year from their commercial publishing ‘partners’ and are completely ignoring where from and how this money was generated. It’s well known that the academic publishing oligopoly is siphoning huge margins of money away from research. Why are academic societies so willingly complicit in this racket? It seems to me as if it is a sadly common approach to deal with this impropriety by turning a blind eye: “Take the money, don’t ask questions!” As long as society members benefit (at the expense of the rest of the world), anything goes.

Some final questions…

  1. Does ESA know how much Wiley is charging libraries around the world for subscriptions to ESA’s journals?
  2. Does ESA actually know anything of the real cost of production and publishing services that Wiley provides – not the price Wiley says it costs (inc. unhealthy profit margin) but the actual cost?
  3. How many readers like me (‘the scholarly poor’) outside the paywall has Wiley charged for access to ESA data papers that should have been free to access?
  4. Given Wiley’s lack of transparency, can we trust them when they report back how many others have also bought access to these ESA data papers that should have been free?

Update 2016/04/09: Thankfully, I did eventually get a refund for this article purchase on 2016/04/08, although I still appear to have lost out due to currency conversion issues with my bank: