Volume VI, 2008
Received 19 November 2007, Revisions requested 4 January 2008, accepted 28 January 2008.
Improving Availability of Print Collections: A Case Study at the College-Conservatory of Music Library
David Sandor, University of Cincinnati
The availability of print materials is an important issue that takes one form in a proliferation of items lost and missing from library collections. The service impact is a matter of direct concern to staff responsible for circulation. This essay discusses availability issues in the collections of the College-Conservatory of Music Library (CCM Library) at the University of Cincinnati. It takes the form of an ongoing case study, i.e. a description and evaluation of management responses to a specific set of availability problems.
What is Availability?
Availability has been a matter of concern in library literature. In his book, Book Availability and the Library User (1975), Michael K. Buckland postulated “immediate availability” as the appropriate standard for library service; ideally, every borrower should be able to obtain every book they want, when they want it (p. 61). Buckland defined the issue as one of physical access to library collections, and theorized that librarians could discover why items were not on the shelf when they were needed and modify library practices to remove barriers to access. Buckland’s approach encompassed collection development and collection management policies as well as circulation policies and procedures; for example, it included user requests for titles a library did not own as well as those on loan or lost or missing from the stacks. Paul Kantor (1976) developed a statistical methodology that measures availability as a percentage of books found from the total number requested. Kantor identified four categories of errors that lead to non-fulfillment of requests: books not acquired, books in circulation, library errors and user errors. In Kantor’s “branching diagram technique” (p. 316), the multiplication of availability rates found for the separate categories produces a rate for an entire collection.
Kantor’s work provided a tool for research. Researchers typically have used questionnaires to discover which books and how many books library users were looking for and how many they found. The questionnaires were analyzed to identify the sources of error for items not found. Policy or procedural adjustments could then be made to improve the availability rate. Examples of such studies include those by Stuart Kolner and Eric Welch (1985), Haseeb Rashid, (1990), and Eugene Mitchell and Marie Radford (1994). Other researchers have proposed the use of sampling techniques instead of questionnaires, in order to simplify the research methodology (Stelk and Lancaster, 1990; Jacobs and Young, 1995; Norton, Seaman and Sprankle, 1996). Thomas E. Nisonger (2007) discovered 50 availability tests documented in 35 studies published from 1980 through 2001.
This case study originated as a practical response to service problems, not a research project. The following account is therefore descriptive rather than systematic. It applies to a subset of availability issues, namely items found to be missing, lost or long overdue during circulation and stacks management activities at the CCM Library. Library staff adopted new procedures to respond to reports of missing items and initiated special projects to isolate and correct errors in the stacks and the library’s database. The goal was to improve service by increasing the availability of items requested by borrowers.
The CCM Library is part of the University Libraries system at the University of Cincinnati. The library supports the academic programs of the university’s College-Conservatory of Music (CCM). The College-Conservatory of Music is the third largest music school in the United States, with more than 1400 students and 170 faculty, and offers degrees in 8 undergraduate and 16 graduate programs through the doctoral level. The CCM Library supports these programs with a collection of nearly 70,000 music scores and 40,000 books and serials held in the Albino Gorno Memorial Music Library, and 54,000 sound recordings held in the Music Listening Center. In the academic year 2006-2007 the University Libraries database recorded more than 51,000 circulation transactions, mostly for books and scores; sound recordings circulate to faculty and staff only. Scores and books circulate to the 86 libraries in the statewide OhioLINK consortium as well as the students, faculty and staff of the University of Cincinnati and some community borrowers. The Gorno Library and the Listening Center are adjacent, but each has its own service desk. The discussion here concerns the print materials that circulate from the Gorno Library. There will be special attention to scores because of the distinctive characteristics of this type of library material.
The impetus came in November 2005, when CCM faculty registered a complaint about lost, missing and long overdue items in the CCM Library print collections. The quantity of missing items accumulated over many years had become an annoyance, sufficient to induce faculty to raise the issue with librarians at a meeting of the Library Committee.
The substance of the complaint and subsequent experience responding to it raised several matters of concern to the CCM Library staff. First, the impact of missing items is disruptive for circulation activities and service delivery. The crisis mode of trying to find missing items wastes the time of borrowers and staff and distracts them from other necessary work.
Second, the complaint indicated that the library might be falling short from a minimal level of service. People who use the library have a right to expect that a score or book listed as available in the catalog will be on the shelf ready to borrow. Materials that are lost or missing should be found if possible or replaced if need be, not left a long time with their status unresolved and their records lingering in the database.
Third, the large number of lost and missing items could potentially amount to significant gaps in the library’s collections. Such gaps might not become known until a book or score was needed, perhaps by a student preparing a recital or a faculty member requesting course reserves. Needless to say these are the worst times to have something come up missing. From a circulation point of view, this amounts to an aggravated form of the first and second problems stated above. In addition, the money and effort spent building collections are less effective if materials acquired to support learning and instruction cannot be found when they are needed.
Finally, these deficiencies if left uncorrected could create an impression of careless, inefficient and wasteful operations, especially for the most observant borrowers. Unless the library made a constructive response, students and faculty could be encouraged to adopt low expectations of library services and collections. This is the opposite of what we should want at a time when borrowers increasingly place a high value on speed and convenience.
An encounter with a member of the CCM faculty illustrates the results of neglecting such issues. Around the time when the faculty were making their complaint to the Library Committee, one professor was unable to find a score she needed for her class and was reporting the problem to the Circulation Supervisor. After the Circulation Supervisor had improvised a problem-solving approach and promised to let her know the result, the faculty member asked, “Do you want to know about this?” Hearing this question from a professor, one can guess that many other students and faculty would have the same question, although they might not choose to verbalize it.
In response to the faculty concerns, the Head Librarian and Circulation Supervisor at the CCM Library established new procedures. The key decision was a commitment to respond immediately to reports of missing items. When a person cannot find a score or book that is supposed to be available, a student employee at the circulation desk verifies the location and status in the database and checks to see whether the material is on one of the re-shelving carts behind the desk. If these steps fail, a student employee or the Circulation Supervisor goes to the stacks immediately to see whether the item can be found there. If it still does not turn up, a student or the supervisor makes a written report with the identifying data for the item and the contact data for the borrower, and searches again later the same day or the following day.
This procedure varies when there is only one person on duty who must staff the desk and is not able to go to the stacks immediately. In this case the written report becomes the trigger for the initial search in the stacks. The library’s commitment is to notify the person about the outcome, whether the item is found or not, as soon as possible but in no case later than seven days. In practice, we notify nearly everyone within one or two days. Our searches locate about fifty percent of the books and scores reported missing on the written reports.
This system is workable in a relatively small library that has an adequate staff. Even so, it is debatable whether such urgency and close attention to the details of missing items are worth the amount of staff time and the priority of an immediate search. The need to put aside other work to look for a missing book or score is not always convenient for staff or efficient for the library, and when the library is busy, if more than the usual number of such reports is occurring can even contribute to an atmosphere of crisis organizing, or “putting out fires.” On the other hand, if the book or score is found the most urgent problem is solved and the user’s need is met. Even if it is not found, the borrower knows the effort was made and if necessary staff can declare the item missing so the borrower can order from OhioLINK. The most important thing in the short term is that something marked “available” in our catalog should either be really available for use, or if it cannot be found an explanation should be made and an alternative provided, if there is one. Missing items deserve priority because the library’s response affects the awareness of the library user directly at the point of service. In this sense, the problem of missing items is not routine.
In addition to initiating a procedure for immediate searching, the Circulation Supervisor also reestablished the routine of making a monthly list of missing items. Each month the Circulation Supervisor or student employees search the stacks for items reported missing during the previous three months. Most of the materials that make it onto this list have already been looked for at least twice before they were declared missing, so the rate of return is not very high. However, we do find some. Generally this list has between 30 and 40 titles.
The establishment of regular procedures set in motion a systematic approach to finding missing books and scores, but as we followed the routines we soon discovered additional problems. First was the discovery of numerous phantom records. These originated from deficiencies in the cataloging module when the database was created in 1991. Phantom records typically bear a creation date from the inception of the database and have no circulation activity and no bar codes. Many are for items that were already missing before the online catalog was created, or else they are duplicate “copy one” records. Because these records appear in the public catalog, they are just waiting for someone to request them in order to become “Missing.”
Examination of these phantom records disclosed another problem. It turned out that many records lacked bar codes because the physical items had never been bar coded. Often these were seldom or never used materials that escaped the original bar coding project when the database was created and were not caught later because they had never come to the circulation desk. We needed to find and bar code these items in order to differentiate their records from the phantom records.
Another question concerns the parts that belong with some music scores. A part is music for one instrument or one voice of an ensemble, usually printed separately from a score and/or from other parts. For example, the music for a string quartet might consist of four parts, or a score and three parts. A score is sewn into a binding, but the parts usually remain separate and are kept in a pocket on the inside back cover. Study or performance of the work requires that all the pieces be kept together. The library facilitates this need by putting a pop-up message in the item record, instructing the person at the circulation desk to make sure all the parts are present when the score is checked out or checked in. The inside back cover is also stamped to show the correct number of parts. At the CCM Library, many item records lacked the pop-up message and some scores were not stamped. Partly as a result of these defects, as well as the normal hazards of use, many scores were missing one or more parts.
The origin of these problems lay mostly in the lack of routine, periodic circulation procedures. It is a question of quality control. The CCM Library does most of its ordering, cataloging and processing in-house. The quality standards of the technical services work are excellent and the errors are few compared with the volume of materials. However, even a few errors per year will accumulate to a large number over a course of years if they are not caught and corrected periodically. The lack of quality control has allowed time for thousands of errors to go undetected. In order to find and correct these errors, library staff initiated several stacks management projects.
We decided to begin by doing an inventory. According to long-time staff, there has been no physical inventory of the collection for about twenty years. Before the era of computerized catalogs, inventory was done manually by shutting down the library for a week or two in the summer and mobilizing the staff to compare items on the shelf with a printed shelf list. The modern method utilizes a hand-held scanner to read barcodes and generate a report from the database. The report indicates the status of each item scanned, either OK or some error status, as well as items misshelved. We began with the reference collection and then moved to the oversize area, with the idea that because these were relatively small we could do them as a trial run and get through them quickly.
The inventory turned out to be useful for both these areas. We found many reference books mistakenly shelved in the stacks and others that belonged in the stacks but had item records still located in reference. In addition there were many mislabeled as well as a few dozen missing and some that had no bibliographic records. Many of the discrepancies resulted from a project done years earlier to reassign materials from reference to the stacks; the books had been moved, but some records were not edited to show the new location. By the time we finished we had a clearer idea of the scope, status and condition of our reference collection of about 4,000 books. For the oversize area, which has about 1100 items, the main problem was incorrect item record locations. Scores and books that are kept in the oversize shelves had item records located in the stacks. Once the records were edited, the reports of items missing from the stacks and found to be in the oversize area declined almost to zero. We finished these two inventories in about two months.
For the stacks, inventory was not as useful as we thought it would be. The most frequent error by far was misshelved items, but these we can deal with more effectively by shelf reading. What is the point of taking something off the shelf, scanning it, then putting it back in the same wrong location? Of course we could correct these problems as we go along, but then we would be doing more shelf reading than inventory. The second most frequent error was the lack of a bar code. This we could resolve more easily by finding the records without bar codes, pulling the items from the shelves and inserting the bar codes. In fact, we launched a bar coding project simultaneously with the stacks inventory, and once the inventory moved into the areas already bar-coded, the number of errors other than misshelved items declined sharply. For the most part there were two or three missing books or scores among several hundred on each report. We therefore assigned a lower priority to the inventory temporarily and prioritized the bar coding project instead.
Bar coding turned out to be relatively easy. It was a matter of running a list, pulling the materials from the stacks, and training student employees to edit the item records. In nine months we pulled approximately 2073 books and scores from the stacks and edited their records. The last phase, now underway, is to run a final list, search again and mark the ones not found as “Missing.” Many of these are the phantom records, which we will delete later with a global edit.
The parts project has been the most complex and time consuming. Each score that needs a pop-up message must be examined and those that lack parts or are excessively worn, soiled or damaged must be set aside. Staff must then do the work of withdrawing most of these and ordering replacements, or sending some to repair. For those not stamped with the number of parts, staff verifies the number in the database and applies the stamp. Student employees do most of the work pulling the scores from the shelves and editing the item records; so far we have edited 3143 records. The initial results of this project have been useful for finding music we need to replace. We have often found that if the score in hand is defective, other editions, copies or volumes of the same title also are. We have replaced many defective scores. In one case we discovered that all six of our editions of the Devil’s Trill Sonata, a violin piece by Tartini, were lacking the part. In some cases parts had been missing as long ago as 1988, according to notes written on date due slips.
Fortunately, the launching of these projects coincided with a recruitment of labor. In spring 2006, the library administration decided to add an extra student worker to the crew during evening and weekend hours. Formerly there was one student assigned in the Gorno Library and one in the Music Listening Center, but now there are two students working in the Gorno Library and one in the Listening Center. This has added fifteen to twenty hours per week to the schedule for special projects in addition to shelving, shelf-reading and circulation duties. The help of student assistants has been critical in carrying out this work.
None of these projects have been completed in nearly two years since we began. It takes time to catch up a backlog of problems after fifteen years. For example, although we have bar-coded more than 2000 items, there are still 1066 item records without barcodes. We are making a second search to find the materials that were overlooked the first time. So far the rate of return on this second search has been about ten percent. For all the items not found, staff must go into the database and change the status to “Missing.” Paradoxically, one short-term result of the work done so far will be to increase the number of missing items by at least 900.
Summary and Conclusions
Lost and missing materials are a perennial problem that requires an active management response. The experience of doing special projects at the CCM Library has highlighted several factors in the management of print collections. One is the need to perform periodic quality control routines. The time spent running lists and tracking down lost and missing items as well as items in transit and other problem categories is necessary for collection management and circulation control.
Next, inventories are an essential management tool. Staffing and budget constraints may make it prohibitive to conduct inventories on a regular basis. An inventory gives only an imperfect picture of a collection, especially if it stretches out over a long period, because collections are always changing. In spite of these limitations, an inventory enables a systematic approach to uncovering problems and can provide an approximation of what a library actually owns. The CCM Library will give it a high priority once the more pressing problems are resolved.
Finally, there is the importance of shelf reading. No amount of slicing and dicing a database and burrowing in the stacks can be a substitute for having things on the shelves in the correct order. We are experimenting with methods to assign and implement a higher priority for this necessary chore.
Of course the goal is not just to have better routines or procedures, but also to have fewer items missing and greater convenience for library users and staff. The results to date have produced raw numbers of errors corrected but no empirical demonstration of a service improvement. The CCM Library still receives about the same number of reports of missing items as we did two years ago. On the other hand, there has been some favorable comment on the fact that the library is replacing scores that have parts missing. This may be an indication that we are on the right track.
The recovery of long overdue materials needs more resources than a relatively small library such as the CCM Library can draw from its staff. University Libraries at the University of Cincinnati has formed a Lost, Missing and Long-Overdue Task Force whose mandate includes the resolution of this category of difficult problems.
The system-wide initiative in the University of Cincinnati Libraries indicates that the type of problems described here are not peculiar to the College-Conservatory of Music Library. One might even guess that they are rather common in many libraries. For example, in 2007 the Office of the Inspector General at the Library of Congress reported that 12.7% of the items searched in the LC stacks could not be found, even after a second search conducted by a Quality Assurance Team. Evidently the problem of lost and missing books needs more attention than it has been getting lately.
Thanks to Mark Palkovic, Head Librarian at the CCM Library, for his comments on this article.
Buckland, M. (1975). Book Availability and the Library User. New York: Pergamon Press.
Jacobs, N. and Young, R. (1995). “Measuring Book Availability in an Academic Library: a Methodological Comparison.” Journal of Documentation 51(3), 281-90.
Kantor, P. (1976). “Availability Analysis.” Journal of the American Society for Information Science 27(6), 311-19.
Kolner, S. and Welch, E. (1985). “The Book Availability Study as an Objective Measure of Performance in a Health Sciences Library.” Bulletin of the Medical Library Association 73(2), 121-131.
Library of Congress (2007). Library Services: Survey of Collections Access, Loan and Management Division Service. Audit Survey Report No. 2007-PA-101/March 2007.
Mitchell, E. and Radford, M. (1994). “Book Availability: Academic Library Assessment.” College and Research Libraries 55(1), 47-55.
Nisonger, T. (2007). “A Review and Analysis of Library Availability Studies.” Library Resources and Technical Services 51(1), 30-49.
Norton, M., Seaman, S. and Sprankle, M. (1996). “Measuring Book Availability: A Monthly Sampling Method.” College & Undergraduate Libraries 3(1), 101-15.
Rashid, H. (1990). “Book Availability as a Performance Measure of a Library: An Analysis of the Effectiveness of a Health Sciences Library.” Journal of the American Society for Information Science, 41(7), 501-07.
Stelk, R. and Lancaster, F. (1990). “The Use of Shelflist Samples in Studies of Book Availability.” Collection Management 13(4), 19-24.