Wellcome Trust APCs: Towards a New [Open Access] Serials Crisis?

In an attempt to make the debate around the costs of open access publishing more evidence based”, the Readme file (14th March 2013) signed by Robert Kiley says, the Wellcome Trust released into the public domain a dataset including details of its open access spend in 2012-2013, “as reported by UK institutions and the Trust’s Major Overseas Programmes in receipt of an OA block grant“.

As I wrote yesterday,

Cameron Neylon subsequently shared a dataset on figshare (and github) with some of the inconsistencies refined:

Neylon, Cameron (2014): Wellcome Trust Article Processing Charges by Article 2012/13. figshare.
http://dx.doi.org/10.6084/m9.figshare.964812

The Wellcome Trust datatset only includes information when an APC was levied. It includes a column for the name of the publisher, as reported by the institution. As the Wellcome Trust does not impose any name authority control on this field, the same publisher was listed in different ways, including typos, acronyms, joint publisher names, etc.  For example, there would be OUP, Oxford University Press, and O.U.P appearing and counting as different publishers. These publisher name inconsistencies were still present in Cameron Neylon’s version of the data as cited above.

I wanted to focus on a few major publishers, and in order to filter them I had to refine the Publisher names inconsistencies a bit. I worked with Cameron Neylon’s version of the dataset and manually refined inconsistencies in the Publisher field (same publishers appeared under different names and spellings and other text formatting issues). I did not refine the journal titles.

I have been looking at that spreadsheet, which reflects the manual refining of the Publisher field I did. As I also tried to explain in my previous post, this ‘refining’ is the result of a human interpretive process, and some of the publisher names that are distinct in that dataset might still be potentially subsumable to other publisher names in the set. Logically, the number of publishers and costs and outputs associated to each publisher will depend on how the Publisher field has been refined; other quantifications and visualisations of the original dataset or other versions that have been refined differently are therefore likely to differ.

After refining the number of publisher names to 101, I  focused on 11 publishers from the dataset and obtained totals as well as their maximum and minimum APCs.

I shared this version of the spreadsheet as

Priego, Ernesto; Neylon, Cameron (2014): Wellcome Trust APC spend (2012-13) Spreadsheet with Publisher Names Refined. figshare. http://dx.doi.org/10.6084/m9.figshare.966427

I am interested in focusing our attention on the highest and lowest APCs that these 11 publishers levied. I believe they offer a glimpse of the average cost of “Open Access” as currently charged by major academic/scientific publishers. I use scare quotes because most of these publishers (if not all?) do not generally publish born-Open Access journals but so-called “hybrid” journals– that is, traditional subscription-based journals that permit authors –ideally via their funders– to pay a fee to make their article available “Open Access”. [Disclaimer: someone still needs to go journal by journal in the dataset to determine as fact which ones are hybrid journals; I haven’t done this yet].

Different publishers call this option differently (for example Springer’s Open Choice or Taylor & Francis’s Open Select). Whether all these “Open” options clearly offer open licensing allowing not only access in terms of viewing but in terms of reuse still needs to be investigated thoroughly.

As shared in my previous post, the following bar chart visualises the lowest and highest APCs levied by these 11 major publishers (click to enlarge).

Lowest and Highest Article Processing Charges from 11 Selected Publishers as Paid by the Wellcome Trust According to 2012_13 Dataset . Chart by Ernesto Priego
Lowest and Highest Article Processing Charges from 11 Selected Publishers as Paid by the Wellcome Trust According to 2012_13 Dataset . Chart by Ernesto Priego

Let’s look at the data table, including the corresponding number of published outputs:

Lowest and Highest APC paid by Wellcome Trust 2012/13 from 11 Publishers, including number of outputs
Lowest and Highest APC paid by Wellcome Trust 2012/13 from 11 Publishers, including number of outputs

[Please note that as explained above these figures are indicative and it is possible that actual numbers vary under a different refining of the Publisher name field.]

Why do I think it’s important to focus on these figures?

For at least two main reasons:

  1. To create awareness through evidence of the price scale of the “Open Access” options offered by hybrid journals from major publishers as paid by the Wellcome Trust (a forward-thinking institution pioneering in their support of Open Access; for their OA policy, go here).
  2. To create awareness of the prevalence of at least three of the publishers, indicating that many scientists still favour them with their work.

It is a truism that “Open Access” was developed in part as a response to “the serials crisis” (on the term, see for example Panitch and Michalak 2005). Major or “legacy” publishers that traditionally have based their business model on institutional subscriptions (toll or paywall) have reacted to Open Access government and institutional mandates by offering “Open” options through Article Processing Charges.

However, these figures reveal what to me at least appears as a mere inversion of the business model, reliant on academic outputs for which considerable funding and/or financial means seems to be taken for granted. The high prices charged to libraries in the paywalled model seem to have been shifted now to the researchers through, ideally, their funding agencies.

It is very important these observations are not misinterpreted as a knee-jerk reaction against all APCs. I edit a journal that charges an APC (and offers its complete waiving as well). Publishing costs money. Enabling Open Access costs money. But does it cost as much as reflected by the APCs in the Wellcome Trust dataset? That is the question.

On the one hand I hope having some awareness of the current hybrid journal APCs charged by major traditional publishers helps provide a point of reference where to judge the current APCs charged by born-Open Access, researcher-led journals like the ones published by Ubiquity Press and other innovative publishers. On the other hand, I believe it is time for those of us involved in enabling Open Access to refine our critical engagement with the term and the current publishing landscape.

The average of all APCs (excluding the £13,000 one for a Palgrave book) in the Wellcome Trust dataset is £1820.01. There is an APC payment for what appears as a single article of £6000. If only all research funders were like the Wellcome Trust. With these rates, who is being excluded from Open Access publishing as currently implemented by the major publishers in scientific/academic publishing? Arts and Humanities research cannot possibly compete. Aren’t we clearly rushing towards a new “OA serials crisis”, where publishing is still dominated by the same major publishers who partly led to the serials crisis in the first place?

Many more questions remain to be asked. Let’s start with those above.

Quick Insights from the Wellcome Trust Paid APCs 2012/13 Dataset

[Updated. I replaced the spreadsheet on figshare twice as a couple of publisher names had to be corrected. This left a version with 101 unique publisher names –note that some might still be subsumable to other publisher names in the set.

I have also corrected the first bar chart and added another one two on this post. Please bear in mind there might still be errors in the source data. The spreadsheet, write-up and charts are shared “as is” and “as available”; the information presented reflects the data as manually curated and refined in the latest dataset version at http://dx.doi.org/10.6084/m9.figshare.966427.

This means that the number of publishers and costs and outputs associated to each publisher will depend on how the Publisher field has been refined; other quantifications and visualisations of the original dataset or other versions that have been refined differently are therefore likely to differ. I do not work for nor am I currently associated with the Wellcome Trust or any of the publishers mentioned here; these are not “official” figures and are openly shared here as research work in progress and should therefore be taken in that spirit].

In March 2014 the Wellcome Trust released a dataset via figshare giving information on their funding of Article Processing Charges in 2012/13.

The dataset  includes all papers that the Trust is aware of paying money for.

Cameron Neylon subsequently shared a dataset on figshare (and github) with some of the inconsistencies refined:

Neylon, Cameron (2014): Wellcome Trust Article Processing Charges by Article 2012/13. figshare.
http://dx.doi.org/10.6084/m9.figshare.964812

I worked with his version of the  dataset and manually refined inconsistencies in the Publisher field (same publishers appeared under different names and spellings and other text formatting issues). I did not refine the journal titles.

I also focused on 11 publishers from the dataset and obtained totals as well as their maximum and minimum APCs.

I shared this version of the spreadsheet as

Priego, Ernesto; Neylon, Cameron (2014): Wellcome Trust APC spend (2012-13) Spreadsheet with Publisher Names Refined. figshare.
http://dx.doi.org/10.6084/m9.figshare.966427

Some figures that stand out:

  • Number of Publishers (once refined): 101
  • Number of Published Outputs in the dataset: 2127
  • Total amount paid in APCs according to the dataset: £3,884,787.52
  • Highest APC in the dataset: £13,200.00, for the monograph: ‘Fungal Disease in Britain and the United States 1850-2000’ (Palgrave Macmillan)
  • Highest APC payment for an article in the dataset: £6000 for ‘Laboratory Science in Tropical Medicine’, in the Public Service Review journal.
  • Lowest APC in the dataset: £45.94  for the journal article ‘The association between breastfeeding and HIV on postpartum maternal weight changes over 24 months in rural South Africa’  on the American Society for Nutrition Journal.
  • APC average (excluding the £13,000 one for the book) £1820.01
Total Cost (£) Paid by the Wellcome Trust to 11 Selected Publishers  (out of 101 Publishers in complete dataset) . Chart by Ernesto Priego
Total Cost (£) Paid by the Wellcome Trust to 11 Selected Publishers
(out of 101 Publishers in complete dataset) . Chart by Ernesto Priego
Number of Published Outputs Paid by the Wellcome Trust  from 11 Selected Publishers in 2012/13 Chart by Ernesto Priego
Number of Published Outputs Paid by the Wellcome Trust
from 11 Selected Publishers in 2012/13
Chart by Ernesto Priego
Lowest and Highest Article Processing Charges from 11 Selected Publishers as Paid by the Wellcome Trust According to 2012_13 Dataset . Chart by Ernesto Priego
Lowest and Highest Article Processing Charges from 11 Selected Publishers as Paid by the Wellcome Trust According to 2012_13 Dataset . Chart by Ernesto Priego

 

Total Cost (£) charged to Wellcome by Selected Publishers  (inc VAT when charged). Figures from 2012/13 Wellcome Trust Dataset (public domain), including Number of Outputs, Max APC and Min APC. From spreadsheet by Ernesto Priego
Total Cost (£) charged to Wellcome by Selected Publishers (inc VAT when charged). Figures from 2012/13 Wellcome Trust Dataset (public domain), including Number of Outputs, Max APC and Min APC. From spreadsheet by Ernesto Priego

With many thanks to Cameron Neylon.

Hopefully this helps in some way to provide a quicker idea of the average cost of APCs from the major for-profit publishers.