Show me the data!

Checking Wellcome Trust funded Open Access articles

February 7th, 2017 | Posted by rmounce in Continuous Monitoring | Paywall Watch

Yesterday I published a blog post calling for ongoing monitoring of ‘hybrid’ open access articles and academic publisher services in general.

Today I want to share with you some highlights from my brief checks on 2 years worth of Wellcome Trust ‘open access’ article processing charge (APC) supported published research outputs.

Source data

Robert Kiley of the Wellcome Trust has made public official data on the APC spend on ‘open access’ articles paid for by the Wellcome Trust over at his figshare profile. This was a brilliant thing to do. Many people have made thought-provoking and brilliant analysis of this data. The data has been copied many times and I see it now in many different github repositories.

Yesterday, just to test the idea, with no real knowledge that I’d actually find anything of interest, I decided to check the DOIs of 2 years worth (2012 to 2014) of Wellcome Trust funded ‘open access’ articles. Here’s the three major things that I have discovered from this mini exercise so far:


1.) Paywalled Articles That Should Be Open Access Wellcome Trust Funded articles that are not openly accessible (updated for accuracy 2017-02-27)

According to Robert Kiley’s figshare data for 2013-2014 Wellcome Trust paid £1,194 to Emerald to make an article entitled Running a hospital patient safety campaign: a qualitative study open access at the publisher website. I followed the DOI link given and found that today this article is paywalled, and is being advertised for sale at £20 for 30 days of access by Emerald Group Publishing (screenshot below):


Sadly, I am no stranger to this kind of event. I have personally seen Elsevier, Wiley, Springer and Oxford University Press sell articles that had been paid-for specifically by funders to be open access to everyone in the world, not to be sold at the point of availability. It seems inevitable now that hybrid open access would lead to this. Paywall publishers simply can’t keep the paywalls off, even if they are paid to do so.

UPDATE 2017-02-27: Three weeks after publication of this blog post Wellcome Trust and Emerald kindly confirmed to me that no APC was actually paid for the above article (contrary to what was mistakenly stated in the figshare data). The article authors backed-out of choosing gold open access. Unfortunately, the authors did not self-archive a freely available version of this paper either, so it remains not freely accessible to those outside paywalls and thus was definitely NOT published in a manner compliant with Wellcome Trust rules and regulations that were in place at the time.


2.) Misuse of funds set aside to cover Open Access charges

I thought this was another simple case of hybrid open access being ‘mistakenly’ paywalled by the publisher but the truth is even stranger.

I found this article entitled, Mechanisms underlying cortical activity during value-guided choice at the journal Nature Neuroscience, that according to Robert Kiley’s data, the sum of £1,272.86 had been paid by Wellcome Trust to make this article open access. The plot thickens however. Nature Neuroscience doesn’t really do hybrid open access, and if it did, it would charge a lot more than that. What I have discovered here is an instance where the authors (or their institution) have mistakenly (fraudulently?) depending on how forgivingly you view this: used the Wellcome Trust Open Access fund to pay Nature Neuroscience £1,272.86 for colour figures. The article is NOT open access at the publisher website. The 2017 Wellcome Trust guidelines are absolutely clear that you cannot use Wellcome Trust Open Access money for “page charges” or “colour figure” charges. I do not know if Wellcome’s rules were so clear back in 2012 when the payment was made but this is extremely disappointing to observe. Charity funding should be better spent than on spurious publisher-invented ransoms like “colour figure charges”.

3.) Elsevier ‘open access’ articles are not accessible to all machine methods

To help check article DOIs in a simple and automated manner, I used the R package httr. The code I used is available as a github gist. The code works well for articles hosted at all publishers except one: Elsevier. Any attempt to follow DOI links with R::httr just hangs and I have to use a timeout to ensure that my script skips over such problems to proceed onto the next article.

Can Elsevier really call what they are offering ‘open access’ if it is not openly accessible by automated methods such as R::httr scripts? I don’t have time to expound upon this at length here, but I will certainly return to this particular point at a later date.


So there you have it. Super simple automated checks of just a few thousand Wellcome Trust funded ‘open access’ articles by their DOIs has revealed three rather interesting things, and supports my overall thesis that we need to continuously monitor academic publishers: not just “one-time” compliance checks.

I really do think this is the start of something very interesting. I have plans. WATCH THIS SPACE!

  • Stephen Eglen

    Hi Ross, how many articles did you find under point 1?

    • So far I have found 5. I have been nice and informed Wellcome Trust about them all behind the scenes. I do stress this is “so far” though. I have been doing ridiculously simple regular expression -based checks on article landing pages of the form grep -L -i ‘Creative Commons|open access’ * which provides a good candidate list to investigate first. I have no doubt that the more I look the more I will find.

  • Roddy MacLeod

    I have seen a number of journals that have an OA version (usually using OJS) and a version within a host aggregator that is behind a paywall and asks for payment for the same articles that are freely available on the OJS site. Next time I find one I’ll try to remember to post here. JournalTOCs always tries to link to the OA versions.