The price of information: Evaluating big deal journal bundles

Increasingly, our access to scientific information is becoming a research topic in itself. Thus an analysis of big deal journal bundles[1] has attracted much interesting commentary (including one from a large scientific publisher[2]). In the UK, our funding councils have been pro-active in promoting the so-called GOLD publishing model, where the authors (aided by grants from their own institution or others) pay the perpetual up-front publication costs (more precisely the costs demanded by the publishers, which is not necessarily the same thing) so that their article is removed from the normal subscription pay wall erected by the publisher and becomes accessible to anyone. As the proportion of GOLD content increases, it was anticipated (hoped?) that the costs of accessing the remaining non-GOLD articles via a pay-walled subscription would decrease.

But as was shown[1], the publishers have hitherto arranged for the prices of these subscriptions to be covered by non-disclosure clauses. Which makes it quite difficult for us (the readers of these journals, and of course the main sources of their content as well) to find out if this model is (starting) to actually work. Certainly, the entire system does not yet appear to be in any sort of steady state equilibrium; perhaps it never will achieve this in the current model? For example, although extra funds have been made available to promote GOLD publishing, these cover only a small fraction of the total output of a typical research university. One could respond to this in several ways:

  1. Find the missing funds from somewhere else, which probably means less money for the research itself. This of course is the model that maintains or increases a publisher’s incomes.
  2. Decrease the costs of GOLD publishing. Currently a typical article processing charge ranges from £500-5000 depending on the prestige of the journal. Is it beyond the realm of possibility that this range could change to eg £50-500?
  3. Simply persuade everyone to publish less. Perhaps ten times less? Every group might be restricted to one or two block-buster articles a year, and the rest of their output goes into open repositories? Or indeed into blogs! These two options of course are unlikely to increase publishers’ incomes.

Well, after 350 years of scientific publishing, we appear to have arrived at a critical point. A cross-roads if you like. But who should be in charge of deciding what direction is now taken? Should it not be the very people who create and then “consume” scientific information and knowledge!

References

  1. T.C. Bergstrom, P.N. Courant, R.P. McAfee, and M.A. Williams, "Evaluating big deal journal bundles", Proceedings of the National Academy of Sciences, vol. 111, pp. 9425-9430, 2014. http://dx.doi.org/10.1073/pnas.1403006111
  2. C. Woolston, "Secret publishing deals exposed", Nature, vol. 510, pp. 447-447, 2014. http://dx.doi.org/10.1038/510447f

Tags: , , , ,

8 Responses to “The price of information: Evaluating big deal journal bundles”

  1. ana lobo says:

    Stands to reason that the producers and the consumers of scientific information should have a determinant say in this matter. If it is not going that way is due to lack of organisation of such lot of people.

  2. Henry Rzepa says:

    Most academics (who produce most of the content and account for much of the consumption) are unlikely to try to exert their influence by e.g. withholding their content. Their prestige, promotion, and future ability to raise grants are far too important to take such a career-threatening action! How else might one account for the observation that they donate their content for free (nay pay up to £5000 to be able to donate their content for free, giving away copyright to boot) and they then donate their time for free (in the peer review processes).

  3. Henry Rzepa says:

    Following up the very last point I made in the preceding comment, here is a positive suggestion. For each refereeing request made by a publisher, the mandatory “payment” could take the form of an APC (article processing charge) coupon, which could be redeemed when the referee turns author and submits their own article. Perhaps three coupons for one APC (on the grounds that a typical article requires three referees). This directly recognises the substantial value that the referee can add to the process.

    You might ask then where is the profit/incentive for the publisher in such a financially neutral arrangement? One could invert this question by asking what further unique value beyond the anonymous peer review a publisher (however you define the term) is best placed to add to the process. Prestige of the publication? But is that the only way to inject prestige? I do feel we must start asking such questions; preserving the status quo is not the way forward.

  4. Darren says:

    Declare a non-reference date: the date after which, papers not published in an open source journal will not be referenced.

    Further, if a reference is regarded as data then the Amsterdam manifesto says that we should be doing this already.

  5. Rzepa Henry says:

    Re Darren. Data publication (as opposed to more traditional narrative publication) is still very much an evolving genre. Its electronic incarnation started off about 30 years ago as supporting data or supporting information (ESI) and most journals tend to make it openly available from the article/narrative landing page (i.e. one can get it even if you do not have access to the subscription-firewalled main article). But ESI as traditionally cast has some serious problems. The documents themselves render the data within almost useless for reuse (often one has to have access to the main article to cross reference against schemes, figures, numbered structures etc), and SI can be almost impossible to mine automatically; humans need to be part of the process.

    More recently, digital repositories have started to be used for ESI or data, and many of the undesirable characteristics of ESI are solved using such access. But I would estimate that greater than 99.9% of currently published articles/narratives do not make use of such enhanced access to data/ESI. In fact, I suspect that most instructions to authors make little or no mention of this possibility. Since by default digital repositories are fully open, increasing adoption of them is surely to be greatly encouraged.

  6. Rzepa Henry says:

    Re Darren. Data publication (as opposed to more traditional narrative publication) is still very much an evolving genre. Its electronic incarnation started off about 30 years ago as supporting data or supporting information (ESI) and most journals tend to make it openly available from the article/narrative landing page (i.e. one can get it even if you do not have access to the subscription-firewalled main article). But ESI as traditionally cast has some serious problems. The documents themselves render the data within almost useless for reuse (often one has to have access to the main article to cross reference against schemes, figures, numbered structures etc), and SI can be almost impossible to mine automatically; humans need to be part of the process.

    More recently, digital repositories have started to be used for ESI or data, and many of the undesirable characteristics of ESI are solved using such access. But I would estimate that greater than 99.9% of currently published articles/narratives do not make use of such enhanced access to data/ESI. In fact, I suspect that most instructions to authors make little or no mention of this possibility. Since by default digital repositories are fully open, increasing adoption of them is surely to be greatly encouraged.

  7. Henry Rzepa says:

    Re Darren. Data publication (as opposed to more traditional narrative publication) is still very much an evolving genre. Its electronic incarnation started off about 30 years ago as supporting data or supporting information (ESI) and most journals tend to make it openly available from the article/narrative landing page (i.e. one can get it even if you do not have access to the subscription-firewalled main article). But ESI as traditionally cast has some serious problems. The documents themselves render the data within almost useless for reuse (often one has to have access to the main article to cross reference against schemes, figures, numbered structures etc), and SI can be almost impossible to mine automatically; humans need to be part of the process.

    More recently, digital repositories have started to be used for ESI or data, and many of the undesirable characteristics of ESI are solved using such access. But I would estimate that greater than 99.9% of currently published articles/narratives do not make use of such enhanced access to data/ESI. In fact, I suspect that most instructions to authors make little or no mention of this possibility. Since by default digital repositories are fully open, increasing adoption of them is surely to be greatly encouraged.

  8. Henry Rzepa says:

    I cam across this detailed analysis of the cost of OA publication: doi: 10.15200/winn.143497.72670 which puts the cost at between $70-600. The lower figure can be achieved if the submitting authors provide XML directly to the publisher, the latter factors in the costs of eg processing Word documents and a whole bunch else.

    In the most popular chemistry journals, the costs seem to be around $1500 or so, and so this figure does seem to include a very healthy profit margin for the publisher.

Leave a Reply