From 1995 to 2008, library-item circulation per student fell 50% at the median across 71 U.S. research libraries for which data are available.[1] This change can be roughly interpreted as meaning that, from 1995 to 2008, students substituted online resources for 50% of their prior university library borrowings.
University library circulation per student declined more rapidly than an earlier decline in public library circulation per library user. From 1856 to 1978, U.S. public library circulation per user was roughly constant. From 1978 to 2004, public library circulation per library user fell 50%.[2] That decrease across twenty-six years is the same magnitude as the decrease in university library circulation per student across thirteen, more recent years.
Younger persons have generally been early adopters of new information and communication technologies. Universities, on the other hand, have not moved aggressively towards online learning. Being lectured and taking notes is still a central feature of formal education. Students are probably leading the way in shifting learning media at universities.
* * * * *
Notes:
[1] For discussion of this trend, see Rick Anderson, Print on the Margins: Circulation Trends in Major Research Libraries, Library Journal, June 1, 2011. Anderson has commendably made the data from his study freely available on the Internet. I calculated the above figure from Anderson’s Table 1 (Word document), with libraries not having data back to 1995 excluded. Data on U.S. university research libraries are available back to 1907. Library data hero Robert E. Molyneux has shared a well-documented, unified collection of such statistics from 1907/08 to 1961/62 (Gerould Statistics). The Association of Research Libraries (ARL) shares online yearly collections of research-library statistics from 1969 to 2008/09. These statistics are not, however, unified and most are not machine-calculable. A machine-readable dataset of these statistics spanning 1988-89 to 2007-08 apparently is available. See ARL Statistics Guide (MS Word document). Analyzing long-term changes in university library statistics would best be done in conjunction with analysis of changes in university enrollments and curricula.
[2] Galbi, Douglas. 2007. Book Circulation Per U.S. Public Library User Since 1856, at p. 2.
I think there’s another way to explain the drop in per-user circulation in public libraries in the 1978-2004 period. This was a time when public libraries made a concerted effort to sign people up for cards. As a result, there’s a much larger, but much less dedicated, user base. Given the fact that some libraries don’t purge their user files even when someone dies (sort of like the Chicago Board of Elections), the number of users has gone up a lot faster than circulation, which has also increased but nowhere near as quickly.
Based on data in my paper on long-term library circulation trends, my estimates of the share of library users in population served are 30% in 1955 and 1978, and 56% in 2004. So those figures are consistent with your explanation.
I wish I had made more library history data readily available. Too bad I wrote that paper pre-online spreadsheets. Now it’s a lot of work to dig out the data.