Mirosław Górny Time as a quantitative measure of the quality of library services
	The Adam Mickiewicz University in Poznań, Department of Linguistics, Institute of Social Communication
	This article was translated thanks to the grant received from the Open Society Institute

Introduction

Time is a fundamental criterion for the evaluation of the quality of information system services. It is primarily important for the user. First of all it is the matter of how long it takes the user to retrieve information from the system, and how much time he or she can devote to processing the information provided. The elapsed time between the appearance of the publication and the moment it is made available is also important for the user. This criterion is crucial for the system administrator as well. He does his best to carry out all the operations taking place in an information system in the shortest possible time (although aiming at shortening the time of individual operations at all cost, in no case can be a rule for the administrator).

This paper will attempt to prove that the indicators used for measuring the time some information system processes take are one of the most important performance indicators of such an information system as library. In some situations these indicators are sufficient for the evaluation of library performance. No matter how they are set, time indicators let us evaluate the quality of almost all aspects of functioning of the information system better than any other indicators. Besides, their one more advantage is that they enable us to use quantitative measures, and thus obtain verifiable and comparable measurement results. The optimisation of the values of time indicators affects the improvement of the quality of all library services.

What indicators to use?

It depends on the abilities and needs of a given library. From among hundreds of indicators used for evaluating the effectiveness and the quality of functioning of libraries we can select a dozen or so indicators for assessing the time of chosen processes; for example the ones listed by Library Performance Indicators ISO 11620 ^[1] are:

1. Median time of document retrieval from closed stacks.
2. Median time of document retrieval from open stacks.
3. Speed of interlibrary lending.
4. Median time of document acquisition.
5. Median time of document processing.

A European Commission study published in 1995^[2] lists the following indicators:

1. Speed of acquisitions and processing.
2. Speed of delivery from closed access.
3. Overall speed in satisfying "requests".
4. Speed of obtaining recalled or reserved items.

The indicators listed above relate to specific library operations, and not the functioning of the library as a whole. That is why some problems arise when using these indicators for the evaluation of hybrid libraries (in the case of electronic libraries there are some attempts at creating separate indicators, for example the Equinox project).

On the other hand, today's libraries are information systems comprising different resources in their functional, technical, and organisational aspects. This in turn makes it possible to devise separate indicators. This paper will focus on only a few general indicators, which seem to be easier to apply than numerous specific indicators, although there are certainly not any objections to using also the latter ones. The topics discussed will be:

• speed of information delivery to the library user;
• the amount of time the user can devote to process the information provided;
• delay - the elapsed time between the appearance of the publication and the moment it is made available;
• the amount of time the user devotes to using the library resources.

The time devoted to information provision

The time of information provision is the period between placing a request for a source of information and making it available to the user. There are two such situations:

1. The user places a request, either by means of a request form or via a network, and after some time he or she receives a print source.
2. The user sends a request (for example he clicks on the source description which he receives as the result of a search) and after some time he is given access to the source material on his PC, palmtop, e-book, etc.

The situation when the user can exactly define the time of placing an order for a source material a description of which has been found in the catalogue, and the time of receiving this material is clear enough. However, we have to examine a few more complex situations. For example the user gets an online access to the source material, and the service he is using has been interrupted (it may be because of technical reasons). Then we propose that these breaks be counted as the time of information provision.

The next situation is when a user places a request for an item which is on loan to another borrower. The time the user has to wait for this item (if he decides on waiting) should be also included when counting the delivery time for the material. Finally, there may be a situation when a library does not hold a requested item and decides on buying it. The elapsed time between placing an order and making the ordered item available should be treated as a time of delivery of this item. Thus we can include the above mentioned indicator called "delay" in the general indicator. We do not take into account here the time of document retrieval (for example searching the catalogue, browsing through Internet resources, looking for an item on bookshelves in the reading room or in open stacks), because, to a large extent, it depends on the user's ability, knowledge and experience. We evaluate only the time which is dependent on the efficiency of the information system.

Provided we have made a proper choice of test options (for example we assign, at random, appropriate indicators to all kinds of user requests related to the access to resources, and then, after these requests have been satisfied, they are being collected for statistics) as the result we get a sequence of data that allow us to assess the speed in satisfying requests and the percentage of requests satisfied in a time period. After all, it is not important for the user how we retrieve the documents requested by him; what really counts is the speed of document supply. Such treatment of the indicator does not require defining the way the materials are obtained, and at the same time it enables us to evaluate the efficiency of the document supply system. If this indicator pertains to some longer period of time, for example a quarter of a year or a year, it will also be possible to take into consideration periodical rise and fall in the number of users' requests.

The indicator for the speed of information provision may be the following: For the time period t the speed of providing information sources was: x1 % of requests, satisfied in the time period t1, x2 % of requests in the time period t2..., xn % of requests in the time period tn.

The set of results that constitute the indicator for the speed of information provision may provide the basis for further analysis. If we find out that one of the values of the indicator differs from the established standards, we should seek the reason for it. If we acknowledge that the value of the indicator complies with standards, we do not need to take any actions.

In practice, it means that we do not take any measures connected with shortening the time period for the supply of information sources, until we do not notice any significantly prolonged time period for the supply for a considerable group of requests. Therefore, the speed of processing documents and adding them to stock is not of interest to us, unless it is not the material purchased in response to the user's order and it does not concern any adequately numerous group of results that are not in compliance with the standard.

The indicator discussed above relates to the time all the processes that apply to providing information sources last. Throughput time for acquisitions and processing depends on the size and structure of library collections (given there is a sufficient number of requested items, the time period for providing information sources is being shortened, since one does not need to wait until the borrowed items are checked in, or to purchase them, or acquire in some other way).

The amount of time spent on using information sources

The user usually wants to have access to sources of information for as long time as possible. In the case of print sources (although not only) the time period has to be limited, at least with a view to the remaining users' interests. The perfect situation would be when the user makes use of information sources for exactly the time period allotted to this type of material (we can limit this time period, for example according to the volume of a given item), but it is also possible to make use of them for a longer period of time if no other user demands access. In practice, libraries prefer to set the same loan period for all items because of difficulties with sending overdue notices. Since this is the time period specified by library regulations, it is not being measured. It is possible, however, to measure the relation between the "usage time" and the information "provision time". In a given situation, the shortening of the "usage time" may mean the shortening of the "provision time". However, depending on the size and structure of library collections, there may be no, or very little, interdependence between these two indicators. The number and the structure of users' requests may also affect this interdependence. The collection features, however, are decisive for the final result.

The amount of time the user spends in the library

Time used in the library is dependent on the speed of providing information. The longer the user has to wait for information, the more time he or she has to spend in the library. Usually the amount of time used in the library depends on the efficiency of the whole library system, for example the quality of catalogues, although, as suggested before, the searching and browsing processes depend largely on the user' characteristics. We may assume, however, that good search tools, organisational efficiency, and the size of a library matched to the number of user requests for information substantially contribute to shortening the time the user spends in the library. It is essential to specify in what way the library's features affect the amount of time used in the library.

Conclusions

Measuring the time is only one of the methods of assessing the quality of library services. It is, however, a considerably easy, cheap, reliable, and universal instrument. If appropriately used and developed, it may very well serve as a preliminary diagnosis of library systems. In order for this to happen, it is necessary to examine what value ranges for the above mentioned indicators and what kind of relations between them point to a specific state of the information system.

Finally, a certain thing that seems to be essential when using time indicators is worth pointing out. Thus, information resources provided by a library that differ in the way they are being used may be divided into two groups. The first group are the sources that are being requested only a few or a dozen or so times a year, but generally users want to make use of them for a longer period of time (they may need these sources for a few months, but each time they study them for a few hours). These are usually monographs, belles-lettres, textbooks, etc. The second group are the sources that users need to study frequently (for example a few dozen or a few hundred times a year), but for a short time (for example a dozen or so minutes each time). These may be dictionaries, encyclopaedias, textbooks, etc. In the case of the first group the time the materials are made available may be longer, but at the same time the library has to guarantee that the user will have the possibility of using these sources for a long time. In the case of the second group of sources the time they are made available to users has to be as short as possible (the usage time will be short considering the kind of information contained in the source). In practice, libraries very often use this division by forming separate collections: the main collection (in closed stacks) and reference collections (for example in reading rooms). For economic and organisational reasons, setting different values for evaluation parameters related to both types of collections seems to be justifiable.

Footnotes

[1]  International Organization for Standardization - Information and documentation: library performance indicators. ISO 1998 (ISO Standard 11620: 1998 E).

[2]  Ward, S., Sumsion, J., Fuegi, D., Bloor I.: Library performance indicators and library management tools. European Commission, DG XIII-E3 1995, pp. 112-113.

Translated by Michalina Byra