AARL

Measuring in-house use of print serials - the University of Melbourne's experience

Helen Thomson

Abstract: This article focuses on the methodology used by the University of Melbourne Library to carry out a survey of in-house use of print serials at the Giblin Economics and Commerce Library in 2000. The article fills a gap in the existing literature in regard to surveys or measurements of the usage of print serials. It is intended in this article to focus on how the University of Melbourne conducted its in-house use survey and so a detailed methodology is included. The results of this survey were then used to identify low use titles in the collection that could be nominated for cancellation. The article also identifies some of the patterns in the results of the in-house use survey and discusses some of the issues related to measuring usage by this method and reviewing print serials subscriptions.

Many academic libraries are faced, almost annually, with the challenge of reviewing their print serials collections. The ever-increasing cost of subscriptions and the roller-coaster value of the Australian dollar has severely impacted on library budgets throughout Australia. Many libraries have adopted a 'zero growth' policy for print serials - in order for a new title to be added to a collection, an existing title with an equal or greater cost must be cancelled. This tough stance has made it imperative that libraries make informed decisions as to the future of their print serials collections. These decisions have been based on the criteria of subscription cost, usage and a ranking of importance or priority given by faculties. While data on the use of e-serials is available of varying quality, the same cannot be said for print serials. In many academic libraries, use of print serials material is limited to inside the library, and as result only limited data are available on the use of individual titles. Collection management librarians have been limited to anecdotal evidence from shelving staff, so collecting reliable data on the usage of individual titles could increase effectiveness in making decisions. To be able to obtain accurate and useful statistics on the level of in-house use of material within an academic library would be invaluable.

Literature review

A literature review revealed that very little has been written on measuring in-house use of print serials and almost nothing on the Australian experience. If similar print serial surveys have been conducted at other institutions the methodology, results and discussion of the issues involved have not yet been published. What has been documented on print serials reviews in the UK and USA has shown that many academic libraries have in the past focused on rankings given by academic staff as the main criteria, along with subscription costs, for either cancelling or taking out a subscription. [1]

Faculty rankings are obtained by surveying the opinion of academic staff who have subject expertise in that particular field. Often these rankings can be influenced by political issues within a faculty so may not always be an accurate representation. Some libraries have made use of citations databases to collect data on the use of print serials by academic staff as reflected in cited journals. Actual in-house use of the print serials collections is rarely comprehensively measured. Detailed methodologies for measuring in-house use of print serials based on a reshelving method have not been documented in the current literature.

Background

The Giblin Economics and Commerce Library is part of the University of Melbourne Information Division. The predominant role of the Giblin Library is to provide information and lending services to the staff and students of the Faculty of Economics and Commerce, as well as the wider academic community. The library includes a monograph collection, a print serials collection, a working papers collection and a collection of Australian Bureau of Statistics publications. These collections total over 60 000 items. The library provides loans of over 70 000 items annually to members of the academic community. The collection has historical importance as it not only includes material from the 20th and 21st centuries but also the 19th century.

The Giblin Library's serials collection includes over 1 800 titles and is heavily used by both undergraduate and postgraduate students as well as academic staff. The collection is available for use only in the library by the majority of library customers. Academic staff and postgraduate research/higher degree students may borrow print serial items for a short term loan of not more than two days. Some of these loans are registered on Innopac - the Library Management System - however the vast majority are recorded as 'manual loans' of which library staff keep statistics. Many print serials items are lent out on manual loan because they are not registered on Innopac as individual items for the purposes of circulation, for example unbound journal issues. As a result little data exists, apart from anecdotal, on the exact usage of individual titles as manual loans only record the number of volumes borrowed and not the titles. The circulation statistics on the library database only provide very limited data as they do not reflect the true usage of the print serials collection.

In 2000 the Giblin Library carried out a survey of in-house use of the print serials collection. The aim of the survey was to see whether accurate and useful data could be collected to identify low use titles to be considered for cancellation as part of the annual print serials review. Data would be collected using a hand held barcode scanner - a barcode would be scanned for each volume used - and then uploaded into Innopac. Once the data were uploaded it could be correlated with the limited circulation statistics of the print serials collection to provide a complete picture of overall usage of the print serials collection.

The survey was conducted in two stages. The first stage involved running a pilot project over three weeks in December 1999 - a relatively quiet period of semester - to test the methodology and feasibility of the project. The results of the pilot project demonstrated that it would be feasible to run the survey. It also became clear from the pilot project that before any title could be judged 'low use' the survey would need to be run during peak semester time to ensure that the data collected accurately reflected use of the print serials collection. A useful side effect of the pilot project was that the data collected not only showed usage of individual titles, but also overall in-house use in terms of the shelving workload for library staff. This provided useful data for rostering shelving staff.

The second stage of the project involved running the survey during semester to collect data reflecting the usage of the collection. The survey was run over 3 weeks during first semester in the lead up to Easter - a time frame deliberately picked as it would coincide with a period of high use by library customers.

Library staff carry out regular pick-ups of library material that has been used in-house in the morning and the afternoon. This material is returned to the sorting shelves so that it can be shelved and easily found by library customers if required. It was decided that the collection of data would be incorporated into these pick-ups. Prior to putting the material onto the sorting shelves the volumes would be counted.

Methodology
The methodology used by the University of Melbourne was modified and adapted from a methodology developed in a similar project by the University of Sydney, in conjunction with the University of Queensland.

Preparation
As the aim of the survey was to identify low use titles to be considered for cancellation only print serials with a current subscription were included. A list of all economics and commerce subscriptions to be included in the survey was compiled using Innopac. One of the initial problems was that not all volumes of a print serial had been added to the library database and given a barcode. To overcome this a dummy record was created for each title that would only be used to record data for the survey. These records were then suppressed from public view on the library's catalogue. The dummy records were organised into broad call number groupings and then sorted alphabetically by title. The records were printed out, a barcode was attached for each title and the printouts organised into a binder that could be easily carried around when data were being collected.

Prior to the survey being run client software for the Percon B321 barcode scanner was loaded onto a PC in the Giblin. Written procedures and training for uploading the data were developed so that the results could be uploaded each day. Procedures were also developed for collecting the data and the staff who were responsible for data collection were also given training in the procedures and using the barcode scanner.

Running the survey
From the results gathered during the trial, it was estimated that running the survey would triple the length of the regular pickups and that additional staffing would be required to carry out the pick up. Casual loans assistants were rostered on each morning and afternoon to collect, scan and then reshelve on the sorting shelves any print serials that were part of the survey. A procedure was drafted and each loans assistant was given training on how to use the barcode scanner.

The assistant doing the pickup and the survey collected, as usual, all material used in the Giblin Library and left at study desks or near photocopiers. The books were unloaded onto the sorting shelves and the journals put to one side. All journals were organised by their call number groupings and volumes of the same title were kept together. Each issue of a title was then counted and the barcode for the title was scanned the appropriate number of times. Bound volumes counted as one item and unbound issues each counted as an item.

Items that were borrowed on a manual loan were scanned as they were returned to the loans desk. It was decided to include this material as, firstly, many academic staff take advantage of this service to photocopy in their own departments, and therefore it is comparable to in-house use; and secondly, this usage would not otherwise be counted on a title-by-title basis.

Uploading the statistics
Each morning, the data from the scanner was uploaded onto Innopac and a statistics report was created. Once the results were uploaded, the previous day's data were erased to avoid duplicating the results. A circulation activity report according to title was produced from Innopac and e-mailed to keep a daily record of the results. A log was also kept of any problems or errors that were discovered during the upload.

Results
Over the three week period that the survey was run 2,437 print serials items were used in the Giblin Library (or taken on manual loan - which in this survey were included with in-house use). This is 17 times the number of Innopac loans for print serials during this period. When the results were correlated it was decided to also include additional information with the results to assist with nominating titles for possible cancellation. This included:

• the current price of the annual subscription for each title
• the priority given by academic staff members of the Faculty in the print serials review in 1999
• the number of times a volume of a title was borrowed via Innopac during the survey
• the total number of times volumes of each title had been borrowed since Innopac was introduced in 1996, and
• whether or not a particular title was available in full text electronically - as this would have implications for the level of usage of the print copy.

At the time this survey was conducted statistics were not available for the usage of individual titles of the library's electronic journals.

The most used title was the Journal of Finance with 138 uses. The top 10 journals according to in-house use were:

Figure 1
Top 10 journals according to in-house use

Although these journals had high in-house use they had less than five checkouts during the three weeks the survey was run. This probably reflects the high use by undergraduate students and the fact that various assignments were due at about the same time as the survey was conducted. It is also worth noting that many academic staff and postgraduate students in the faculty are taking advantage of increasing availability of full text journals from their own offices via the Buddy network - the University of Melbourne's gateway to the library's electronic resources - or else make use of the short term loan system. The short term loan system allows academic staff to leave their staff card in exchange for a journal to quickly photocopy in their own department free of charge. These short term loan statistics were included in the internal use figures.

Faculty rankings were included for comparison purposes. It is interesting to compare what users are using in the library with what the faculty consider to be important for their own teaching and research. Serials are ranked in order of their importance for faculty research and teaching purposes and also overall importance for the subject area. The various departments within the faculty are asked to rank the serials within their subject area and some level of consensus must be reached. In previous print serials reviews, the faculty's rankings and price have been the main criteria for cancellation of a subscription. For example, in 1999 a subscription to the journal, Problems of Economic Transition, cost $1 646.30 and was ranked by the faculty as having marginal importance. This would be an ideal candidate for cancellation. This was further supported by the survey when the journal recorded zero uses during the period the survey was conducted.

In comparison, the top 10 journals according to total circulation were:

Figure 2
Top 10 journals according to total circulation

Total circulation figures refers to the number of times volumes of a print serial title have been borrowed on Innopac since 1996. Figures are relatively low because only staff and postgraduate students can borrow bound volumes. Some of the titles with high circulation figures are treated as monograph in series and are available for loan to undergraduate students. The Accounting and Auditing Handbook has very high checkouts because although its backfile is treated as a serial, multiple copies of the least two to three years are treated as recommended reading monographs. Similarly, there are recommended reading copies of the World Development Report in the monograph collection.

The top 10 journals according to price were:

Figure 3
Top 10 Journals according to Price

There would appear to be little correlation between the cost of a print serial title and its usage. The most highly used titles were not always the most expensive titles. Results of the survey indicate that 75% of print serial titles with an annual subscription of A$1 000 or more recorded low levels of in-house use. Figure 1 which lists the top journals according to in-house use indicates that only one journal title in that list had a subscription cost of more than A$1 000. The most used title - Journal of Finance - had a relatively inexpensive subscription cost of A$373 pa.

More than 200 titles recorded zero internal use. Of these 25% were ranked by the faculty as essential. Almost 150 titles recorded low internal use, ie 1-5 uses, and 33% of those titles were ranked by the faculty as essential. There were 52 titles with 6-10 uses during the survey, while 34 titles recorded more than 10 uses. Of those titles that were used more than 10 times, 66% were considered by the faculty to be essential. Thirty-two titles (including the top 10 titles) recorded 20 or more uses and 84% of those titles were ranked as essential.

In almost all of the categories listed above the top 10 journals were available electronically but usage of the print volume continued to be high. Although no concrete data exist as to why this is the case possible reasons may be either that users were unaware that the title was available electronically or did not know how to access it; users were unable to access a terminal to gain access to the online version; or users perceived it to be quicker and easier to access the journal in print rather than electronically.

Issues

Difference between In-house use & faculty rankings
The relationship between the actual use of individual titles and the faculty's opinion of individual titles - expressed as a ranking - is a key issue. The opinion of academic staff has in the past been a key factor in whether titles have been cancelled or subscriptions maintained. These opinions are very subjective. Key staff representative of an academic department's teaching and researching interests are asked to indicate which journals are either essential, desirable or marginal to their subject. Often those titles marked as essential match the top peer reviewed/refereed journals in a subject area. Decisions to mark a journal as essential, desirable or marginal can be the result of internal politics within an academic department or by the research or teaching needs of an individual academic. These rankings can also influenced by the absence or presence of a key staff member, an example of this would be when a staff member is on sabbatical and returns to find that a important journal for their subject area was cancelled.

Relationship between cost of a journal, use of a journal and faculty ranking of a journal in maintaining or cancelling a subscription.
The three criteria - subscription cost, usage and faculty ranking - are crucial in determining whether or not subscriptions should be maintained or cancelled. Previously, many libraries have relied heavily on the cost and faculty ranking of a journal as criteria because data on usage was not obtainable. As it becomes more and more difficult to identify titles for culling - all obvious candidates having been cancelled in previous rounds of print serials reviews - usage begins to play a more important role in the decision. It provides a different perspective on a print serials collection, and while usage alone is not enough to justify cancelling a title, used in conjunction with cost and faculty ranking it can help build a more complete picture.

Reliability of in-house use data
While the survey has provided useful data on the in-house use of the Giblin print serials collection, the reliability of the data collected needs to be considered. The data were collected over a very limited period of time. Usage of the collection was undoubtedly influenced by the assignments and subjects being taught at the time. Running the survey during another semester when different subjects were offered and different assignments are being completed would possibly produce quite different results. As a result of the restructure and staff changes the survey was not repeated. It had been intended to conduct the survey twice - first during a similar period and second at a later date in the academic year. This would have provided additional data and allowed for comparisons to be made. It was hoped that by repeating the survey the exact impact on subject based use could be determined.

Other factors that could influence the validity of data include whether or not library staff are collecting the data and how often data are collected. In the survey, data were collected in the morning and the afternoon. Volumes were counted as they were taken from the sorting shelves to be reshelved. It is possible that some volumes were not included as library customers reshelved their own journals. It is also worth considering that volumes may have been used more than once in the time that they were waiting to be counted. In this instance, a volume may have only been counted for a single use rather than a possible two or three uses. The impact of electronic access also needs to be considered.

Impact of electronic availability
Another issue that needs to be considered in light of the growing crisis concerning print serials collections in many libraries is the issue of electronic access. In December 2000, 24% of the University of Melbourne Library's serials collection was only available electronically, ie there is no corresponding print subscription. The majority of titles in many disciplines are still only available in print. In addition, many print subscriptions are joint electronic/print subscriptions which include access to the online version of the print title as long as a print subscription is maintained. It is still valid to consider the importance of in-house use because the majority of use is still via print sources where available. Statistics on the proportion of the library's print serials collection that are also available electronically are not currently available.

One set of data that were not included in the survey which needs to be considered in a print serials review is that of the use of print journals in electronic format. Increasingly, libraries are relying on electronic journals to fill gaps and are also making cancellation decisions based on the availability of a particular title in full text. In some cases, libraries hold an electronic subscription to a print serial title in conjunction with a print subscription. Under these circumstances to get an accurate picture of usage of that title both print and electronic usage must be counted - as many library customers are taking advantage of the convenience of online access to avoid the pitfalls of using the print version in the library. Statistics for electronic usage of the print serials included in the survey were not available as it was only in hindsight that the importance of such data was realised. In future surveys, arrangements should be made to collect data on the availability and usage of electronic journals to be included with the results for an comprehensive view of the serials collection.

Validity of faculty rankings
As the factors that can influence the validity of usage statistics need to be carefully considered so too those factors that influence rankings given by academic staff. These include individual bias; whether the staff member in question has a research or teaching role and is therefore making decisions for the benefit of their own research or whether a print serial title might be required by undergraduate students. The weighting given to the faculty ranking is also important. If a single staff member conducts research in a subject area for which there is only one key title is his or her 'essential' ranking more valid than for an area in which several staff members are conducting research but in which there are also several key journals considered 'essential'. In this case, faculty politics can be an important factor.

Disadvantages of this method
A major issue - and one that will need to be considered by any library interested in carrying out a similar survey - was that of staffing. It would not have been possible to conduct this survey without additional staffing. The survey doubled the staff time required to carry out regular pick-ups of material used in the library. An extra nine hours a week (total 27 hours - @ HEW 2) was required to carry out the scanning. The amount of staff time needed to upload the statistics and correlate the data also needed to be taken into account. In total 11 hours were spent uploading the statistics each morning and more than two days were spent collating the results and compiling this report.

Conclusion

The survey was a success. It has provided useful data on in-house use in the Giblin Library. Previously, the only data available on in-house use were anecdotal, and that usually limited to generalised statements. The in-house use survey provided a title by title breakdown of in-house use for most of the titles for which the Giblin Library has a current subscription, albeit for one relatively short period. By repeating the survey's snapshot of the collection's use at other times throughout the academic year, it would be possible to build a more complete picture of in-house use and allow comparison.

It is, however, unfortunate that due to a restructure of the University of Melbourne Library and resulting changes in staff and stakeholders in the project that there was not more follow up conducted on the use of this data in the print serials review. The data were used to identify some low use, expensive titles with a marginal faculty ranking that were cancelled in the 2000 print serials review. Further follow-up, ie conducting the survey in other semesters to collect more data has also not been possible. However, the survey which was conducted as trial to test whether or not this methodology was a feasible way of collecting usage statistics that could be used in print serials reviews was successful. The methodology has proven sound and that these type of statistics in combination with loans statistics, price of subscriptions, faculty assigned priorities and electronic availability and usage would provide a basis for a systematic evaluation of the use of currently subscribed print serials held in the library.

Notes

1. K E Joswick and J K Stierman (1995) 'Perceptions vs. use: comparing faculty evaluations of journal titles with faculty and student usage' Journal of Academic Librarianship vol 21 pp454-8; T W Klassen (2001) 'Measuring serials usage using faculty cited journal data' The Bottom Line: Managing Library Finances vol 14 no 1 pp12-19; L J Sweeney (1999) 'Confronting crisis: the importance of measuring the use of paper journals in academic libraries' New Library World vol 100 pp72-79

Helen Thomson, earth sciences librarian, Information Division, University of Melbourne, Victoria 3010. E-mail: helenlt@unimelb.edu.au.nospam (please remove the '.nospam' from the address).