Author’s note: With my colleague Karen Barch and the support of the Andrew W. Mellon Foundation, I have been working on a project to determine how many books university presses publish, and in particular, how many of these could be termed original monographs in the humanities. The full report is available below as a PDF and also on Scribd. The text of this blog post is a slightly edited version of the report’s Introduction. For the quantitative aspect of the study, I refer you to the full report.]
This is a report on a linked series of projects to study the output of American university presses in the time period 2009-2013. A large amount of data was collected from the presses themselves; it was then aggregated and analyzed. This analysis yielded reports for the internal use of the Andrew W. Mellon Foundation (“Mellon”), the sponsor of this research. Subsequently we prepared individual reports for the participating presses. We also experimented on a project with OCLC in which we attempted to map (with only partial results) the database of press titles to the holdings in WorldCat. This report (the “public report”) represents an edited version of the various private reports submitted to Mellon and the presses. The primary thrust of the project can be found in Section V: Phase Two: The Core Database of the full report.
The genesis of this project came about from conversations in 2013-2014 within the academy concerning the condition and future of university press publishing and, in particular, the academic monograph, which is an essential component of scholarly communications and which also plays a role in the credentialing of scholars. There was at that time only anecdotal information on how many monographs university presses published and, hence, no obvious way to measure the size and scope of university presses and the certification system they help to support. (“Do university presses publish too many books? Too few? Do they support some fields more than others?”) In order to get some data to assist in other explorations into the university press world, Mellon asked us (that is, Barch and Esposito) to tabulate the output of university presses; that tabulation lies at the center of this report.
This report represents our attempt to answer Mellon’s question, a three-part project involving American university press book publishing, which we conducted, with some breaks along the way, over three years. The three parts will be taken up in turn in the body of this report; they consist of a pilot project, a study of the output of American university press monographs over the period 2009-2013 (the core of the project), and an experiment in which we (with the assistance of personnel from Mellon and OCLC) attempted to map, with mixed results, the information we gathered in Phase Two onto WorldCat, an OCLC service.
It is worth bearing in mind that university presses publish many things, and monographs are but one slice of their overall production, which could include journals, trade books, tests, regional titles (e.g., a tourist’s guide to the region where a press is located), and classroom texts. Mellon specifically asked us to determine the number of scholarly monographs in the humanities published each year by American university presses. This immediately raises the question of what is a monograph. In our survey of the presses, which were asked to fill out a large template, we used the following instruction:
For a definition of monograph please use John Thompson’s from Books in the Digital Age: “books which are written by scholars and researchers and which are intended primarily for other scholars and researchers” (p. 103). To Thompson’s definition we add the following: for purposes of this study, a monograph can have more than one author, but please exclude as monographs books that are collections of essays, even if the essays are all by a single author.
Another question is whether or not an edition is original. For example, if a press publishes a monograph in a cloth binding one year and a paperback (or an ebook) of the very same book a year later, does that count as one monograph or two? We thus introduced the concept of original works, designated in the survey as “primary” works (primary because most original monographs eventually spawned other editions).
The survey we conducted had a potential sample of 106 presses, 65 of which successfully filled out our template and submitted the data to us. The template the presses used also allowed us to capture some other useful data about the presses — such things, for example, as pricing data, the number of titles published that are not monographs, etc.
The reporting presses published a total of 58,555 books over the five-year period, for an average of 11,711 books per year. The number of these books that were described as primary (that is, original) monographs was 14,619 for an average of 2,924 per year. And the number that was described as primary monographs in the humanities was 10,689, an average of 2,138 per year.
Since not every press filled out the template, we extrapolated what the five-year and one-year results would have been if we had had 100% participation for the presses (exclusive of Oxford, Cambridge, and the Associate Members of the Association of American University Presses, henceforth “AAUP”). The extrapolated figures are:
- Total books: approximately 76,000 or 15,000/year
- Primary monographs: approximately 19,000 or 4,000/year
- Humanities primary monographs: approximately 15,000 or 3,000/year
In the body of this report we introduce greater granularity to this analysis. We have divided the press community into six groups (defined in the report) and provide summaries for each group. We have also extracted data for four subject areas (art, history, literary criticism, and philosophy) in order to assess output at a finer level.
While it is not possible to know if there is a consensus view, we were surprised that the monograph output in the humanities for American presses (exclusive of Cambridge, Oxford, and the Associates) is as small as it is. This suggests to us that programs designed to intervene financially on behalf of monographs will be challenged by the fact that monograph publishing resides in a broader business context, making it difficult to untangle such things as the allocation of overhead.
We do not believe that there is sufficient evidence for the five-year span of this study (which may not be long enough to extrapolate trends) to assert that the level of humanities monograph output has decreased. Our view is that it probably has not, but to prove this point one way or the other would require a study over a longer timeframe. A potential project for 2019 would be to add five years of data to this study (2014-2018), which would enable more meaningful trend analysis.
One item that came up early in our work with the participating presses was that of confidentiality. Many presses expressed discomfort in disclosing so much information about their operations. We suspect as well that since we were collecting information on sales and pricing among other things, some presses were being careful not to leave a suspicion that they were colluding with other presses. Because of this, all the information provided to us has been held in confidence, shared only with Mellon.
We thus prepared private reports to Mellon on the data. The first of these was on the pilot project, the second on the rollout of the full survey. We also drafted another private memorandum on the experiment we conducted with OCLC. Internally, we have referred to the report attached herewith as “the public report.” The public report eliminates certain fields of information and any connection between the data and a specific university press. Neither the private reports nor the underlying dataset will be made public.
While it is not possible to know if there is a consensus view, we were surprised that the monograph output in the humanities for American presses is as small as it is.
A note on the presentation of the data. The data falls into three categories: the raw data, the summarized data, and the data that is inserted inline in the report. The first two categories are private, the third public. But the data is intelligible to anyone with access to the private reports who takes the time to study the column and row heads (it is represented in large spreadsheets). We made some of the summary data available to Mellon and will shortly be sending it to the individual presses that participated in the study, but it is not inserted inline here.
Which brings us to the OCLC experiment. The background on this experiment derives from research performed by Rick Anderson and Dean Blobaum and published on the Scholarly Kitchen. Rick and Dean sought to determine how important library sales are for university press monographs. They took a sampling of titles from The University of Chicago Press, where Dean is employed, and looked up the ISBNs on WorldCat, using WorldCat to identify library holdings. Alan Harvey of Stanford University Press saw this blog post and realized that it would be possible to get a fairly comprehensive picture of the library distribution of university press titles by mapping the data from our project onto WorldCat. We attempted this, but ran into the problem that OCLC does not identify books as publishers do (OCLC is far less reliant on ISBNs). Thus the mapping to WorldCat was only partly successful.
The OCLC experiment raises the question of what other uses the data can be put to. One use was made by Paul Courant and Terry Geitgey, who investigated “free riders” in university press publishing. (A free rider in this context means an author whose parent institution does not have its own university press.) We suspect, though, that the most practical use will be made by the participating presses, to which we will be distributing individualized reports on their own programs and how their programs compare to those of university presses of similar size and to the university press community overall. Those individual reports will remain private.
For readers who wish to get right to the primary findings of this study, please go directly to the report: Section V. Phase Two: The Core Database.