Previous Contents Next
Issues in Science and Technology Librarianship
Summer 2013


Are We Spending Too Much on Print STEM Monographs? A Method and Analysis for Improving Monograph Allocations Based on Circulation Statistics

Jonathan Nabe
Collection Development Librarian
Southern Illinois University Carbondale
Carbondale, Illinois

Copyright 2013, Jonathan Nabe. Used with permission.


Circulation studies provide evidence of demand for monographs, but it is necessary to determine the goal of any analysis in order to select which statistics will be used. The goal of this analysis was to determine the appropriateness of expenditures on monographs within the STEM fields at Morris Library over a ten-year period. Percentage of unique title circulation and average circulation per title are best suited for this purpose. Results show variation among discplines, but over-expenditure in all. Analysis of disciplines can aid in determining appropriate allocations for monographs, and analysis of subdisciplines can aid in targeting monograph acquisitions within any discipline.


Library collection budgets are under pressure from a number of fronts. Declining federal and state support, a long-term problem exacerbated by the recession that began in 2008, have resulted in smaller increases, or in some cases even reductions. Although serial price increases have moderated somewhat in the last few years, they still exceed inflation and exceed increases for most libraries' collection budgets (Bosch and Henderson 2012). These pressures typically affect monograph budgets disproportionately, and force a closer look at how monograph collections are being used, in order to determine the appropriate allocation of funds and to match purchases with patron needs. This paper describes a method and analysis of monograph circulation in the STEM fields at Southern Illinois University Carbondale.

Circulation statistics provide a window into the needs and behavior of users, and can be used to answer a number of questions pertinent to collection management. The first task for a circulation study is to determine the goal of the analysis. This influences what statistics are gathered, how they are arranged, and what analysis is applied to them. Inevitably, the goal of the analysis will determine what methods are applied.

There are (at least) three specific questions which circulation statistics can help answer:

What is the demand for monographs by local patrons at the institution?

Circulation statistics can be used as one part of an analysis of the total (monographic) demand of local users, but need to be coupled with interlibrary loan borrowing data as well as e-book usage data, and perhaps citation data, pulled from the publications of faculty and students (in the latter case, particularly theses and dissertations).

What is the demand for the institutional collection?

Circulation statistics can be used to assess the use of the collection of not only local users, but all users; interlibrary loan lending data will need to be included in this case. This can be useful, among other things, in determining the role of the library in cooperative collection development, and the collection's importance to other users, including (via public libraries) local and regional users not affiliated with the institution.

What is the adequacy of the local collection for local users?

The goal here is to determine the appropriateness of budget expenditures for monographs (and perhaps the quality of book selection). It is this question which formed the goal and guided the method of the following study.

To determine the appropriateness of budget expenditures on monographs for local users, interlibrary loan lending and some in-house use (such as checkouts to preservation) must be excluded from the analysis. If the goal is to reveal local demand, extra-institutional use is irrelevant. Similarly, reserve circulations should be excluded, particularly when a separate reserve fund exists, since their circulation provides a misleading indicator of the collection's appropriateness.

It is important to decide whether to include browsing data in the analysis. Browsing data refers to the tracking of in-house use of collections other than in-house processing events. At SIUC, barcodes from books taken off the shelf but not checked out are scanned; this constitutes a browse which is recorded and attached to the item record, and is available as a separate field in circulation reports. There are a number of problems with such data; first, the percentage of browsed books actually scanned is unknown. Student workers are largely responsible for this function, and compliance is indeterminable. Second, it is unclear what this data signifies. Books are pulled off from the shelf for many reasons, not all of them in furtherance of some academic purpose. Library staff sometimes pull misshelved titles and leave them on a table to be reshelved; these titles may be scanned as a "browse," but drawing any conclusion about their demand from this would be mistaken. It is simply impossible to determine the meaning of a "browse." While the same criticism may be applied to circulated titles - not all of them are read, and many are found to be in some way insufficient - it is known at least that the patron took the trouble to take the book to the circulation desk and check it out. There is this extra step and an assumption of responsibility for the book that make a "charge" qualitatively different from a "browse." Accordingly, browses are not included in this analysis.

Another factor to consider is the time period to be covered, for both the age of the monographs, and the date of the circulation. This, too, depends on the goal of the analysis. When addressing the question of appropriateness of budget expenditures on monographs for local users, only recent data will provide an accurate picture of library use and the adequacy of selection practices. The goal is to equate the demand of current users and present and near-future expenditures; how the collection was used twenty or thirty years ago does not help to meet this goal; it fails to account for changes in curriculum, programs, and enrollments. Similarly misleading is the circulation of legacy collections, those that were acquired before these changes in curriculum took place. This is not to say use of legacy collections and analysis of that use has no value; indeed it is essential in determining the need to maintain those collections. It does not bear on the question at hand, whether current and recent selections correlate with current and recent use, and how to adjust budget allocations accordingly.

Overemphasis on only very recent data is also misleading. Restricting data to a few years fails to account for fluctuations in programs enrollments, which is common in most institutions. The user base for any particular subject or department will oscillate from year to year, so their use of the collection needs to be averaged over some period of time. Current and very recent use fails to capture potential use; not all monographs will circulate immediately. There can be no hard and fast rule, but ten years, for both the age of monographs and date of circulation, is a reasonable compromise.

Once these parameters are established, reports can be run and data compiled. A new set of questions then arises: how is the data to be analyzed? What measurements will provide useful information to answer the question of the appropriateness of the collection and budget expenditures?

Two measurements stand out. The percentage of titles that have circulated at least once reflects the extent of any over-expenditure by revealing the gap between demand (use) and supply (collection). The average number of circulations per title reflects the extent of the demand for monographs as a whole, by accounting for multiple users of all titles.

Together, these two measurements provide an adequate indication of the appropriateness of budget expenditures for monographs. When arranged by Library of Congress Classification, (LC), they provide the basis for comparison by discipline.

This study does not consider e-books, because the multiplicity of models for e-book acquisition complicates analysis. The method described below could also be applied to individually selected e-books, but Morris Library relies on a patron driven acquisition (PDA) model for most of its e-book acquisitions, supplemented by purchases of collections from some publishers. Different models of acquisition require different methods of analysis; this study is restricted to a method designed to provide information about one model of acquisition for one format (print).


SIU-Carbondale is a Carnegie RU/H university located in Carbondale Illinois, and has both a Law School and a Medical School. Enrollment is 18,847, including 4,700 graduate and professional students in 74 masters and 32 doctoral programs. Morris Library is an ARL Library, with over three million volumes and 51,000 serials. The Library has a collection budget of $5.2 million. The Law and Medical Schools have their own libraries and budgets, but there are no other branch libraries.


Circulation statistics for a ten-year period, from 2002 to 2012, were compiled from the statistics reporter of the Voyager system. Only those titles purchased during this time period were used. Results were restricted to items with charges to faculty, students, and staff of the University, by using the Patron Groups field and removing the circulation counts for all other patron groups. This eliminated interlibrary loans and local charges. Data fields included in the report were Create Date, Title, Location, Normalized Call Number, Reserves Status, and Historical Charges. The Reserve Status field was used to eliminate titles that were placed on reserves at some point in the ten-year period. The Location field was used to eliminate titles in reference and other locations that prohibit circulation.

The resulting spreadsheet contained only those titles purchased in the ten-year period that could circulate, had never been on reserve, and had been checked out only to local patrons or had never been checked out.

Separate spreadsheets were created for each call number range at the highest level (for this analysis, Qs, Rs, Ss, and Ts). These were then broken down to match the subjects under review (QC for Physics, QD for Chemistry, etc.). Thus each subject included a list of titles and their historical charges. Analysis was conducted on these subject spreadsheets.

Data calculated included the total number of titles, the total number of charges of those titles, the total number and percentage of titles with at least one charge, and the average number of charges per title for the entire subject.


All disciplines show a percentage of unique title circulation rate below 50%. In other words, for each discipline, fewer than half of the titles purchased in the ten-year period were charged at least once. The total percentage of titles with at least one charge is 38%. Average charges per title were below two across all disciplines, and the overall average charge per title ratio is .86.

There is a distinct difference among the disciplines for both percentage of unique titles with at least one charge and average charges per title. Table 1 shows the breakdown of the nine disciplines. Three disciplines (Computer Science, Medicine, and Math) exceeded the average percentage of unique title circulation rate, and each had a charge per title ratio above 1. All other disciplines show a unique title circulation rate below 40%, and two (Chemistry and Geosciences) are below 25%.

Field # of Titles Total Charges % titles
≥ one charge
Computer Science 1,934 2,788 48%
Medicine 9,416 10,301 45%
Math 3,046 3,169 43%
Engineering 13,265 10,908 37%
Physics and Astronomy 2,616 2,117 36%
Biology 7,747 5,856 36%
Agriculture 4,665 2,895 31%
Chemistry 1,224 617 23%
Geology 1,890 863 23%
Table 1. Circulation statistics for STEM monographs purchased 2002-2012.


The unique title circulation rate is similar to that of other studies, including the University of Pittsburgh study from 1979, a recent study at Cornell, and the overall rate at ARL libraries, where the circulation rates were reported at 60%, 45%, and 44% (Stewart 2011). A five-year study at the University of Nebraska-Lincoln revealed a 54% circulation rate (Tyler et al. 2011). Those studies examined all subjects, not just the STEM fields, and it should not be surprising to find slightly lower numbers in the STEM disciplines. The University of Colorado reported a 33% unique title circulation rate across all disciplines (Knievel 2006). At UNLV, circulation rates for science titles ranged from 14-21% and for engineering from 14-24% over a period of five years; the shorter time period may partially account for the lower numbers there (Tucker 2009).

The difference among STEM disciplines is clearly evident in the unique title circulation rate and the charge per title averages. The advantage of looking at the collection use by call number as opposed to the originating department of the patron is that it incorporates evidence of interdisciplinary demand. What is important is not who checks out monographs, but rather the demand that each discipline exhibits.

Clearly, Agriculture, Chemistry and Geosciences are not disciplines with high levels of monograph use. Their low charge per title ratios show that books as a whole are infrequently charged in these disciplines, and the low unique title circulation rate shows that when they are charged, a small number of titles satisfies demand. In contrast, the charge per title rates in Computer Science, Medicine, and Math (all above 1) indicate that books are in demand, and the higher unique title circulation rate shows that there is more diversity in the demand.

The numbers also reveal that over this ten-year period, monograph acquisitions have far exceeded demand. When budgets are insufficient to meet patron needs, the cost of over-expenditure in one area (i.e., purchasing resources that are not used) are too high and cannot be ignored (Carrigan 1996). As mentioned in the introduction, determination of total demand for monographs would require combining these circulation numbers with data on interlibrary loan borrowing and electronic book use. Nevertheless, on their own they clearly demonstrate that overexpenditure has occurred.

The data revealed not just overall demand, but how that demand is distributed among subdisciplines. Analysis by subdiscipline allows selectors to target those areas where demand is highest within a specific field, thus maximizing increasingly scarce resources.

Table 2 shows a partial breakdown of the same measurements, unique title circulation and charges per title, for a number of subdisciplines within Math. From this it becomes clear that monographs in certain areas, higher in demand, are more appropriate targets for acquisition than others. A study is underway to determine whether selection based on this granular analysis leads to higher performance in terms of the two measurements, higher unique title circulation rates and charges per title.

LC Subject LC Class # Titles # Charges % Titles
≥ one charge
Average charges per title
Analysis General QA300-QA302 36 81 75% 2.25
Linear Algebra and Matrices QA184-QA205 62 112 65% 1.81
Popular QA93-QA99 60 93 63% 1.55
Calculus QA303-QA316 95 182 61% 1.92
Mathematical Statistics QA276-QA280 246 439 59% 1.78
Algebra General QA150-QA161 82 130 59% 1.59
Probabilities QA273-QA274.9 194 257 53% 1.32
Theory of Functions QA331-QA355 70 47 50% 0.67
Continuous Groups QA385-QA387 10 15 50% 1.50
Differential Equations QA370-QA381 141 189 49% 1.34
Combinatorics QA164-QA167.2 124 129 47% 1.04
Analytical Methods QA401-QA402.37 54 43 44% 0.80
Numerical Analysis QA297-QA299 61 69 44% 1.13
General Geometry QA440-QA497 64 80 44% 1.25
Topology QA611-QA614.97 155 151 43% 0.97
Number Theory QA241-QA247.5 118 100 40% 0.85
Functional Analysis QA320-QA329.9 61 51 38% 0.84
Group Theory QA174-QA183 50 41 34% 0.82
Arithmetic QA101-QA146.8 89 61 34% 0.69
Machine Theory QA267-QA268.5 33 16 33% 0.48
Differential Geometry QA641-QA672 52 33 31% 0.63
Math General QA1-QA63 693 443 30% 0.64
Fluid Mechanics QA901-QA930 37 13 30% 0.35
Algebraic Geometry QA564-QA609 45 23 27% 0.51
Analytic Mechanics QA801-QA835 42 25 26% 0.60
Table 2. Circulation Statistics for select Math subdisciplines for monographs purchased 2002-2012.


Responsible budget oversight and expenditure requires data-driven analysis and decision-making. Particularly in times when journal cancellations are a routine part of conducting business in academic libraries, monograph fund allocations need to be consistent with use, and title selection must be targeted as precisely as possible to avoid spending money on unneeded resources. Unique title circulation rates and average charges per title ratios provide clear evidence of monograph demand, and can help guide selection decisions appropriate for the institution.


Bosch, S. and Henderson, K. 2012. Coping with the terrible twins. Library Journal 137(8) 28-32.

Carrigan, D. 1996. Collection development - evaluation. The Journal of Academic Librarianship 22(4): 273-278. DOI: 10.1016/S0099-1333(96)90117-2

Knievel, J.E., Wicht, H. and Connaway, L. 2006. Use of circulation statistics and interlibrary loan data in collection management. College & Research Libraries 67(1): 35-49.

Stewart, C. 2011. METRICS: The next chapter: measuring the pace of change for print monograph collections. Journal of Academic Librarianship 37(4): 355-357. DOI: 10.1016/j.acalib.2011.05.001

Tucker, J. 2009. Collection assessment of monograph purchases at the University of Nevada, Las Vegas Libraries. Collection Management 34(3): 157-181. DOI:10.1080/01462670902962959

Tyler D., Melvin J., Yang X., Epp M., and Kreps A. 2011. Effective selectors? Interlibrary loan patrons as monograph purchasers: a comparative examination of price and circulation-related performance. Journal of Interlibrary Loan, Document Delivery & Electronic Reserves 21(1-2): 57-90. DOI:10.1080/1072303X.2011.557322

Previous Contents Next

W3C 4.0