Loughborough University logo

IFLA Glasgow 2002

IFLA Satellite Conference

Statistics in practice - Measuring & Managing

An official IFLA satellite preconference sponsored by the IFLA Section on Statistics

13-15 August 2002, Loughborough, Leicestershire

Speakers' Abstracts

Statistics is a necessary part of a well balanced professional education

Patricia Layzell Ward

A case for the importance for information professionals to possess a basic understanding of statistics is argued. A study of the teaching of statistics in ILS schools in a number of countries indicates that most schools do not, apparently, now include a course, module, or unit on statistics. The implication of this for the development of future managers and researchers is discussed. The role and requirements of accrediting and professional associations are reviewed. Proposals are outlined for ways to overcome this deficit in preparation for a professional career.

But what does it mean? Using statistical data for decision making in academic libraries

Steve Hiller

Academic libraries are increasingly turning to the use of statistical data to assist and support library management decisions. While some library data, primarily operational, has been available for years, the degree to which it was used in library management varied widely. Recently, there has been a voluminous increase in library-related data, not only in transactional information from online library systems and electronic resources usage, but also from efforts to gain more direct user input through such methods as user surveys and focus groups. Libraries are also being asked by funding and accrediting bodies to demonstrate their impact on the user community by employing a series of metrics that are outcomes based which provides even more data.

As the data flow turns into a flood, many academic libraries are unsure what the data means, how to analyze it, and how to use data for in library management. This presentation focuses on methods to identify, analyze, make sense of, report and use relevant data to improve performance in academic libraries.

Using statistics to enhance library performance, an example

Anja Smit

This paper will discuss the application of statistics in a medium sized academic library, the Universiteitsbibliotheek Nijmegen in The Netherlands. It will focus on the recent implementation of revised statistics and the introduction of national benchmarking. You will be given an example of how statistics can play a growing role in enhancing the performance of an academic library.

Benchmarking for improvement

Liz Hart

This presentation will examine benchmarking in libraries to achieve improvement in services. Libraries can learn a lot from each other and benchmarking provides a structured framework for making comparisons between organisations. This presentation will provide a very brief overview by methodologies before progressing to provide some detailed practical examples of how academic librarians have evaluated and improved their services through benchmarking. There will also be helpful examples of the potential pitfalls involved in the process as well as ways of avoiding them!

From Kaleidoscope to Common Sense

David Lightfoot

Making real use of statistics to achieve service improvement needs to be part of the basic training of library managers at all levels. David will show how Lancashire has embarked on a long term programme to help staff make effective use of statistics. Lancashire has recently commissioned a joint project with LISU and Resource the Advisory Council for Museums Archive and Libraries. David will describe his experience and concerns with statistics in libraries that lead him to embark on this project. He will expose the strengths and weaknesses of some of our current range of statistics and standards in Libraries using a kaleidoscope of examples from his experience as a library manger and in his various national roles. This paper will highlight the need to produce a strategic plan if our considerable investment in statistical collection is to lead to real service improvement.

Measuring What Matters: Organizational Effectiveness by the Numbers in One Canadian Public Library

Don Mills

The presentation will concentrate on a decade of using statistics and measurement methods to manage a large urban public library (serving 650,000 with 16 locations) in a period of rapid growth and development of its services.

The Mississauga Library System doubled in 10 years and yet successfully met all the challenges of growth to remain the top rated local service. Through a formal management process composed of several key activities including performance management, strategic management, organizational health, and value management the Library identified key statistics and key quality as well as quantity performance indicators and sought to affect those positively through changes in inputs and outputs in key areas of service expansion, innovation, continuous improvement, and efficiencies. The presentation will review the components of the Library's formal management process and the measurement methods used as well as a multi-year approach including major service plans and their multi-year budgets. Examples of the statistics collected, the methodologies used, and performance indicators devised for each part of the process will be examined along with the Library's evolving vision for and definition of ultimate "success" - superior service at a reasonable cost.

As well, the key Canadian library statistics - public as well as other - will be reviewed as the background to one large public library's pursuit of effectiveness. The issues of the Canadian library scene will be covered along with current national practices and the unmet measurement needs today.

Putting statistics into practice - strategies for effective management

Eric Davies and Claire Creaser

The aim of this session is to explore, in a practical hands-on context, the range of options for using statististics and performance indicators for effective management.

The objectives will be to:

  • develop an appreciation of the importance of performance indicators and statistics in managing effectively;
  • identify the different aspects of management that can be supported and enhanced by appropriate data;
  • raise awareness of the data that are available; and
  • demonstrate how these can be applied in practice.

    A short introduction to the range of data and techniques which can be utilised for effective management will be followed by a series of practical workshops, where a range of real-life problems will be examined and discussed. There will be a break for coffee, and a general reporting and discussion period at the end.

    Sampling statistics of library use

    Sebastian Mundt

    Whenever it is practically impossible or too costly to analyse the population in total libraries commonly apply sampling procedures to survey users, measure library performance (misshelving, book processing etc.) and evaluate specifics of the collection (percentage of foreign literature, loss of books etc.). The revised International Standard ISO 2789 "International Library Statistics" now allows for the use of sampling procedures to estimate annual totals of library visits, in-house use, information requests and - where necessary and applicable - the use of electronic services.

    Literature on sampling in libraries regularly provides thorough information and guidance on estimating percentages (e.g. in survey sampling). Sampling of library use, however, frequently aiming at total numbers, is often being melt down to the recommendation that it should be "grossed up from normal weeks". In contrast to a random selection of the sample the purposive pre-selection of "normal weeks" implies detailed knowledge about the variable in question.

    It is well-known that, for examples, daily use of academic libraries' services is being influenced by general factors like the "academic year", events inside the library and the availability of "competitive" library services on the campus. It can be argued furthermore that a number of more obscure or "random" factors like technical readiness of buildings and systems, local daily weather conditions or important cultural or other events in the vicinity may have their impact on "physical" library use as well. It is difficult, however, to reduce these often complex practical circumstances to a statistical problem, to quantify its parameters and predict "normal" activity. Some libraries have shown that analysing time-based, local and thematic characteristics of data can at least be useful for grossing up data from a sample. By examples of real data the presentation describes and compares different methods of random and non-random sampling in the library environment, and it evaluates different ways to estimate annual totals from the sample count.

    Electronic journal usage statistics: present practice and future progress

    Tony Kidd

    Comparable and standardised e-journal statistics are desperately needed by library staff, for a number of purposes including collection management, collection promotion, budget allocation, and budget negotiation. This paper reviews the requirement for usage statistics, and then moves on to consider some of the efforts that have been taking place over recent years to promote and implement standards in this area, concentrating on the ICOLC guidelines, and on the international work in progress under the auspices of the PALS Usage Statistics Working Group.

    Measures for electronic use: the ARL e-metrics project

    Julia Blixrud

    Libraries are spending a larger proportion of their materials budget on electronic resources. In 2001, members of the Association of Research Libraries (ARL) spent more than $132 million on electronic resources. In order to determine if their dollars are being well spent, 24 members of ARL self-funded a research project to develop measures for describing the resources, expenditures, and usage of electronic resources. The project resulted in a set of recommended measures and a data collection process that will continue to be refined and tested by ARL members.

    Measurement-based change in libraries: a case study

    Joan Stein

    Practicing librarians are often frustrated by the extent of assessment theory available to them and the lack of easily accessible examples of putting this theory into day-to-day usage in their own libraries. There is a real interest and need within the profession for worked examples of data collection and analysis used to inform management decisions and create positive changes in meeting users' needs. This presentation will address specific examples of using data collected at various service points in an academic library to improve service to the user community. It will largely address regularly and/or easily collected data from a variety of sources instead of special, large-scale studies that require extensive effort and/or specialized knowledge. The advantages and disadvantages of certain data collection techniques will also be discussed.

    Using statistics to allocate the academic library materials budget

    Wanda Dole

    Academic librarians have proposed a variety of methods and formulae for the allocation of library materials budgets by discipline and by format. Recently the need to allocate funds for electronic resources has presented an additional challenge.

    Percentage Based Allocations (PBA), first proposed by David Genaway in 1986*, is a method by which the percentage of the library's budget allocated to each discipline is equal to the percentage of the total university budget for instruction and departmental research received by the corresponding academic department or program. This presentation examines the use of PBA at two academic libraries of very different size and type.

    The presentation describes an eight-year test of PBA at a large research institution. It discusses the rationale for selecting PBA and compares PBA with other allocation methods or formulae. It postulates that PBA is an objective method for the distribution of library materials budgets funds in alignment with university and library priorities. The presentation also examines the question of PBA's scalability to a small and medium-sized institution. The literature does not contain any examples of implementation of PBA other than that at the two institutions studied.

    * David Genaway ,"PBA: Percentage Based Allocation for Acquisitions: A Simplified Method for the Allocation of the Library Materials Budget," Library Acquisitions: Practice and Theory 10 (1986), 287-292.

    Opportunities and Success Stories

    John Sumsion

    In concentrating on current problems for a professional audience we easily overlook what statistics demonstrate in the broader historic and geographic context. Sometimes the success is so complete that NILs (or tiny figures) are now recorded, for instance, for: Closed access material; Material not on the OPAC; Copying done by staff; Charges for book borrowing. Other changes are so huge that they go unrecorded: Items requiring original cataloguing; Recording access (entrance) to the library; Security staff controlling physical access. But inter country comparisons show that such signs of the modern service are not yet universal.

    Examples, from public and academic libraries, are discussed where particularly successful initiatives and projects comprised important statistics in their origination and/or monitoring. References to other speakers' examples are included.

    Statistics are considered that are appropriate for a one way interpretation or for an average rather than an extreme target.

    Contrary to fashionable opinion there is a need for more statistics about library operations rather than less - though their selection for each target audience has to be undertaken sensitively to avoid negative overload. Academic researchers could profitably spend more effort in collating and analysing statistics from individual libraries to monitor current trends and development.

    Careful use of statistics is essential for effective library management. In practice those libraries that provide a superior service have managements attuned to quantitative assessment of their results. Under resourced or poorly managed services are generally weak in their statistics.

    Areas where management is historically weak also feature statistics that are relatively underdeveloped, for example Cataloguing, Cost Accounting.

    The future scenario of networked, generally unmediated services will call for major changes in librarians' statistical armoury with some new performance targets. Work towards these is in progress. Regular statistical series have, by definition, to monitor and follow rather than precede actual development. However, some speculative (and provocative) targets can usefully be formulated at this stage.

    LISU gratefully acknowledge the support and sponsorship of
    Emerald and Swets Blackwell for this event.

    Copyright © LISU. Updated 7 July 2004.