From “What is the ullage of your library?” SPARC Open Access Newsletter, January 2, 2004.
What percentage of the journals or articles needed by the faculty at your university are unavailable except by interlibrary loan or private emails to lucky colleagues elsewhere? Let's call this the “ullage” of the library, after the word for the empty space at the top of a wine bottle. The ullage of a library is the gap between what is directly available and what is needed.
I bring this up, of course, because rising journal prices increase ullage, and spreading open access decreases ullage.
The Los Alamos National Laboratory (LANL) library has made a start in measuring the ullage at 13 major research libraries in the U.S. Instead of looking at all the journals that an institution might need, it looked only at the ISI list of the 100 most-cited journals for 2002. Fair enough, since we can assume that all research institutions would want access to these journals. Instead of counting any kind of subscription, it counted only electronic subscriptions revealed on the library web sites. This seems to be a methodological shortcut to save counting time, but insofar as research universities want electronic access to their most-used titles, it should not distort the measurement.
Kathy Varjabedian published the resulting bar charts on in the December 2003 issue of the LANL Research Library Newsletter.
Bottom line: even the best-stocked research libraries have regrettable ullage. If we look at all the e-journals studied, then the best-stocked library is at the University of Illinois at Champaign-Urbana, which holds 98% of them. For this body of literature, it has an ullage of 2%. If we look only at the journals outside clinical medicine (a fairer measurement for universities without medical schools), the best-stocked library is Princeton's, which holds 70 of the 70 most-cited journals outside medicine. It has an ullage of 0%.
There are two ways to bring ullage to zero. We could provide open access to still-needed resources or find enough money to buy access to them. Because OA isn't the only way to do solve the problem, ullage doesn't measure the progress of OA so much as the itch that OA or money can scratch. We could say that ullage measures the problem, not the solution, but we have to bear in mind that it only measures the reader-side problem (need without access). We need another way to measure the author-side problem (contribution without audience or impact).
Erratum [added in 2004]: I misread the LANL article. Princeton has 70 of the 70 most-cited journals outside medicine, not 70 out of 100. Hence it has a ullage of 0. The figure I cited for the University of Illinois at Champaign-Urbana is correct. I apologize for the error and thank Greg Price for alerting me to it.