Creating (the) Future (for) Libraries
Image by JenWaller via Flickr

When I served as a collection development librarian between 2000 and 2006, I was involved in three sequential journal cancellations.  The first year was easy — cancel paper, duplicates, and obscure foreign titles.  Year Two was more of a challenge.  By Year Three, it was hard work.

This process is still going on, and fortunately, I am no longer part of it.  No one gets thanked for canceling journals.

As part of the program to educate our faculty, we took our message on the road and visited each department in our college.  We went in twos, and I was partnered with the head of access services.  His role? Advertise new services for faculty (extended loan periods, electronic reserves, and book delivery straight to faculty offices!)  My role? Tell them about what we needed to cut and lecture them on the consequence of their publishing behavior on the library.

In order to make my point about the disparity in journal prices, I would bring props to these meetings, illustrating how much content they received from a year’s worth of PNAS, say (I had to use a book cart for this subscription) compared to an expensive commercial niche journal.  The journals would be chosen purposefully because my talk was as much theatrics as it was about message — some journals cost a lot of money; we can no longer subscribe to them all; it’s time to make hard choices.  In retrospect, the message hasn’t changed.

I succeeded in delivering this message to every department I visited, and with each succeeding performance my confidence grew.  I was going to help change scholarly publishing in my college, and help herald in a new era of responsible publishing.  SPARC would be proud of me.

And then I was taken to task.  At one meeting, a scientist stopped me mid-performance.  My comparisons were not only wrong, they were misleading. Some of my journals published many more articles than their competitors; others levied page charges to authors, thus artificially lowering the prices paid by libraries; some offered publications to society members.  Therefore, I should understand that libraries subscribers are only one part — albeit an important one — in the economics of a scholarly journal.

This scientist (renowned in his field) was kind in his delivery, but the message resonated throughout his peers.  I learned that if I were going to convince scientists to change their behavior, I needed to be upfront about the facts I present and about the facts I left out.  Librarians, after all, are supposed to work in the faculty’s best interests, not in their own.

In his piece, The Library: Three Jeremiads,” Harvard University Librarian Robert Darnton mourns the decline of the scholarly press, the spiraling of journal prices, and the tenuous relationship the academic library has entered with the Google Book scanning project, a position Rick Anderson critiqued in a prior post.  For the very same reason that I was critiqued for cherry-picking facts and presenting a narrow view of publishing economics, it’s important to clarify Darnton’s argument about journal prices.

For those of us who have been in the library, publishing, or higher-education profession for even a short period of time, Darnton’s victim narrative is formulaic:

  1. Present the fact that journal prices have been escalating far faster than the general rate of inflation
  2. Create a caricature of all publishers as commercial behemoths concerned only with profit
  3. Select the most-expensive journal as an example of corporate greed (i.e. Brain Research)
  4. Bemoan the library as victim of this relationship
  5. Finally, present open access as a unified solution to the crisis

You’ll find this narrative almost everywhere you look in the library literature, and I am very familiar with it because I used it frequently myself.

In previous posts, I argued that the Consumer Price Index was a very bad comparison for measuring the purchasing power of libraries, that journals are the wrong indicator for growth, and that describing the present situation as a “crisis” is both inaccurate and unhelpful for finding real solutions.

In this post, I will argue that the price of a journal is a very bad indicator of value to an institution.

To emphasize the escalating price of journals, Darnton selects Tetrahedron, although he forgets to alert the reader that, like Brain Research, Tetrahedron is a collection of journals, which includes:

  • Tetrahedron
  • Tetrahedron Letters
  • Tetrahedron: Asymmetry
  • Bioorganic & Medicinal Chemistry
  • Bioorganic & Medicinal Chemistry Letters

And with this combined collection of journals, Darnton also omits the fact that he is quoting a combined subscription price ($39K).  I am not going to defend this price nor the act of offering a bundled subscription — Elsevier can defend themselves.  I do however take issue that Darnton compares this five-part chemistry journal to an average humanities journal.

ISI has not finished indexing the Tetrahedron collection completely for 2010, but these five journals have published nearly 6,000 articles in 176 issues during the past year.  In comparison, most humanities journals publish 16-20 articles in just four quarterly issues.  Divide publication numbers into the yearly subscription price and you get about $6/article for Tetrahedron and between $15-$19/article for the average humanities journal. That expensive behemoth starts looking like a good deal.

Now, I didn’t include the fact that many humanities professors purchase their own print copies of journals, thus inflating the total institutional price of having access to a journal.  Nor did I mention that the library often pays again for online access to humanities journals through bundles like JSTOR or Project MUSE — costs that were not added to Darnton’s figures.  When normalized for the size of these journals, Tetrahedron is no longer the poster child of poor value.

Cost per download analysis often tell the same story: large expensive science journals are cheaper than small humanities journals.

After lamenting the state of the Harvard library budget — arguably the most enviable collections budget in the world — Darnton describes steps that his university has taken to find a new sustainable economic future. One of these solutions is the Compact for Open-Access Publishing (COPE), a compact that “could save billions of dollars in library budgets.” It is not clear, however, if Darnton considers Harvard’s libraries in those that could save.  Of the nearly 6,000 articles in the Tetrahedron bundle, Harvard researchers authored 22  of them in 2010.  Given that COPE will pay $3,000 for each article out of this fund, paying for open access would cost Harvard $66K in 2010, $27K more than its subscription price.

Given the guidelines for the COPE signatories, many of whom refuse to cover publication charges when the author has a grant, libraries could technically save money if they require their faculty to pay more of the freight of scholarly publishing, although this is simply a form of shifting — not diminishing — costs.  And holding library funds in reserve for when authors or their grants cannot pay may be doing more economic harm than good.

The last few years have not been good for academic libraries and I, like most in higher education, feel their pain.  Yet, distorting the facts in order to present the situation in a simplified perpetrator-victim narrative is simply not helpful.  It’s time for the library community to write a new story.

Enhanced by Zemanta
Phil Davis

Phil Davis

Phil Davis is a publishing consultant specializing in the statistical analysis of citation, readership, publication and survey data. He has a Ph.D. in science communication from Cornell University (2010), extensive experience as a science librarian (1995-2006) and was trained as a life scientist.


16 Thoughts on "Poor Comparison Leaves Darnton's Journal Price Jeremiad in Jambles"

This is a marvelous example of how a blog can offset advocacy by presenting additional (that is, omitted) facts. Well done.

I am not at all impressed. His confession that as a librarian he had resorted to questionable information tactics and “coloured” depictions of reality does not give me particular confidence in Mr Davis’ present intellectual credibility. This is not what I have to come to admire in the past as a role model when following U.S. professional librarian’s mailing lists. Mr Davis present critique of Darnton’s piece is “small coin” if compared to Darnton. Davis builds a strawman in order to have something to demolish. Darnton did not compare Tetrahedron to the average humanities journal. In fact, he compared the price of the average physics journal to the average humanities journal. Also he listed such figures just as a rethorical devise to set the scene for his Jeremiad. A Jeremiad doesn’t claim to be a scientific analysis.

If Darnton presents us with a Jeremiad, he obviously has a narrative to tell. Davis has not; he cannot even claim the position of a “minor prophet”. It is unclear what Davis wants to tell us when criticising Darnton that he “forgot” to tell that “Tetrahedron” is a package. It is simply irrelevant, if the figure is anyway just used for illustrative purposes.

In my limited perspective of a large technical university library, in terms of cost per download, JSTOR is much better value than all our STM packages, but this only as an aside.

Following Davis’ logic, we should probably tell our humanities profs: Sorry, but your journals are so expensive in terms of cost per article, that we’ll cancel all of them. The money is much better invested in Science journals where large output allow much higher efficiencies. (Never mind that the cost of humanities journals isn’t the cost driver that brings our libraries into trouble.)

I think Phil makes some very good points, and the comparison with COPE is particularly telling about how much of a bargain
“open access” might really be for a top research university like Harvard. I find the reply above by Mr. Kaemper confusing, especially the comparison in value between STM packages and JSTOR. Isn’t that comparing apples and oranges since JSTOR (until very recently) offered only legacy, not current, content?

I do have one question, though. Given that many science journals have a very high rate of acceptance (over 50%) whereas most humanities journals don’t (10% is typical), can we be sure that we are really getting equivalent “value” from journals in these two areas? It would seem that libraries are being asked to pay for a lot of “C”-grade material in science when they are getting mostly “A”-level material in the humanities, if acceptance rates are a fair criterion of quality.

There may be points to criticize Darton on, but to talk about the value of science journals versus humanities journals is a ridiculous one. The American Historical Review accepted only 8% of articles submitted last year, and received 3,000 books to review (humanities journals aren’t just articles, by the way). Also, humanities journals generally don’t respond to more submissions by creating spin-off journals for B and C publications. They just reject more articles. Science publishers have become very adept at making a value case based on sheer volume, which you unthinkingly buy into. But most scholars (in any field, sciences or humanities) will tell you that it’s quality that counts, not a flood of articles for $39,000.

I think we are getting off track here. The point of my post was not to compare the relative value of the STM versus the humanities literature. It was to point out that that the narrative constructed by Darnton was based on selective data, which, analyzed more fully, provides counter-factual evidence to the position he was attempting to make.

Darnton promotes an old narrative that views the library as a victim of a simple injustice perpetrated by greedy, profit-seeking commercial publisher. I argue that casting oneself as a blameless and innocent victim is not helpful and finish by suggesting that the library needs to find a new narrative — hopefully framed positively — that helps chart the future of academic libraries.

Not all academic librarians support this simple victim narrative (Ivy Anderson promotes a more realistic view that views open access solutions in a more positive light). While her view is much more nuanced and takes paragraphs to explain, I believe it is a more helpful (and hopeful) library narrative to promote.


with respect to JSTOR: We have licensed this as a journal package, and judge its usefulness via cost per download, just as we do for other packages and aggregator databases. By this measure, it is excellent value for money, better then for most STM packages. And it is not comparing apples and oranges. Especially in the humanities, an article which is needed by a scholar is not less useful because it is a few years old. Also, our clients find it extremely useful to have this broad and diversified resource available, even if we could never afford to license current versions of all these journals.

The problem with C grade material in science journals is that it tends to be a lot more expensive then the A grade material because the latter’s circulation is higher. This is the reason why you need to cancel most journals from publishers like Elsevier and Taylor & Francis in order to get affordable and efficient package deals.

While they’re harder to boil down to sound bites, there are several measures of journal cost effectiveness that correct for things like volume of articles and the like. One, for instance, is Eigenfactor’s cost-effectiveness search, where you can see how much libraries are paying for each “eigenfactor” unit (which essentially measures influence by the degree of citations that journal articles get from other influential sources).

In any given field, you can see quite a large spread of cost-effectiveness values. (Comparisons across different fields may be misleading, though, which is one reason this particular search only compares journals in the same general subject area).

The endless argument about whether science or humanities journals are a better “value” is pointless. If your university teaches humanities, you need humanities journals and books. If it does sci/tech, it needs sci/tech journals. This is not a library decision. Ultimately the costs of supporting subject disciplines must lead to universities setting priorities and cutting unproductive or unsustainable programs. Information costs of those programs, onerous as they are from the library budget perspective, are a very small piece of the pie. Unfortunately that piece is usually decoupled from the larger budgetary questions and left to be considered in isolation by librarians who do not have the bigger picture in mind because they are minor functionaries.

Information budgets need to be returned back to their respective disciplines so that the units directly affected by these decisions actually have some say in them and some reason to care. At present they don’t. And thus faculty are subjected to woe-is-us presentations from librarians who would never otherwise darken their doors.

David, I’m not sure where you get the idea that libraries make acquisitions decisions “in isolation” from the “bigger picture” of disciplinary need. The fact that we talk a lot about pricing issues doesn’t mean that we make our acquisition decisions primarily on a cost-per-unit basis. On the contrary: we tend to use local need as our primary criterion when considering a new subscription, and local usage rates (usually measured by downloads) as a primary criterion for retention. (Cost-per-download and cost-per-article-published are very different criteria; the former is locally relevant, while the latter, as you point out, is locally irrelevant.) Any librarian worth his salary will consider an understanding the curricular and research needs of the campus absolutely essential to the responsible performance of his duties as a librarian.

I have heard some librarians say that, although STM journals are generally much more expensive than humanities journals, “political” considerations dictate that when journals are cancelled, humanities journals must be included even though cancelling STM journals saves a lot more money. That is a library decision, though dictated by the realities of campus politics. I daresay many scientists are unaware of this reality, however, or simply don’t care. This is not based on any attempt to measure comparative value, but just on some (perhaps perverse) sense of fairness.

Phil says that page charges to authors and faculty print subscriptions to humanities journals ‘artificially’ lower the prices paid by libraries. I think this comment misses the point of the scholarly communication crisis entirely. Before the commercial sector became a major force in journal publication (and before journals went online), academic journal publishing revenue was far more diversified than it is today. Advertising, memberships, personal subscriptions, and page and other publication charges all played a part in sustaining journal publication. Today, a far greater percentage of the revenue burden has shifted to libraries. Add to that the steady increase in research publication over very long time frames and the equally steady erosion of library purchasing power in higher education when measured as a percentage of institutional expenditures – trends that have been well-documented by Michael Mabe (writing for STM) and ARL respectively for many years – and you have a structural dilemma of crisis proportions that calls for a radical rethinking of the economic arrangements underlying the scholarly publishing enterprise. Assuming that journal publishing rates are not going to decline, the solution would seem to lie in two directions: diversify revenue streams to incorporate additional sources of funding, and lower the unit cost of publication. The former can be accomplished in part through publication-side funding from authors and research funders; the latter seems only likely to be achieved through new publishing initiatives that seek to re-establish publishing using lower-cost methods and infrastructure. If both of these were to be achieved, libraries at research-intensive institutions would arguably be in a better financial position than they are now, and the market would be better balanced. Open access is the publishing paradigm that seems most likely to effect this change. This is of course completely apart from the societal benefits of unfettered access to the research literature. Open access offers the potential (although I would agree with Phil, not the assurance) of advancing both of these goals.

Thanks for this even-handed comment. This is a more nuanced argument that avoids the library-as-victim narrative and projects open access in more of a positive light rather than a negative reactionary one. What is surprising is that the library-as-victim narrative still dominates the message coming from academic libraries — or more specifically, those who speak publicly for academic libraries. In this sense, I think we both agree that the narrative has to change.

I think there is a more interesting story about victims to tell here, and it involves commercial publishers only indirectly. Ultimately, the harm that humanists have suffered from declining purchases of monographs owing to the high costs of STM journal subscriptions was caused by their colleagues in the sciences who, mostly from ignorance or inattention, fostered the rapid growth of STM journal publishing through their own practices (needing, e.g., to have large numbers of published articles on their resumes to be competitive in attracting new research grants) and then brought pressure on their librarians to purchase all these journals. Considering the political power that scientists wield on campuses by virtue of the external funding they bring in, it’s not surprising that librarians acceded to their wishes–and cut their budgets for monograph purchases instead. Commercial publishers merely took advantage of what was offered to them and handsomely profited in the process. But they did not start the process that ended with the resulting harm to humanists.

“Of the nearly 6,000 articles in the Tetrahedron bundle, Harvard researchers authored 22 of them in 2010. Given that COPE will pay $3,000 for each article out of this fund, paying for open access would cost Harvard $66K in 2010, $27K more than its subscription price.”

In an effort to clarify the actual costs to Harvard if the Tetrahedron journals were open access, I reanalyzed the actual situation based on the actual workings of COPE-compliant open-access funds (and Harvard’s HOPE fund in particular). It turns out that the cost to Harvard would be $0, not $66,000. Under other counterfactual assumptions I explore, the cost varies between $6,700 and $32,500 — far less than the current subscription price. I provide full details in my post at my blog.

More importantly, as I argue there, the idea behind COPE and the HOPE fund is not to save an individual institution money, but rather to support a functioning market-based system for scholarly article publishing. If in the long term COPE has the intended effect of shifting journals such as Tetrahedron to an OA model within a journal ecology based on an efficient market, and if under those conditions Harvard ends up paying a bit more, well, then the market has spoken.

Comments are closed.