Collection Level Description

A review of existing practice eLib supporting study

[previous] [next]

3. Collection Description

3.1 Conspectus

3.1.1 Origins

The Conspectus methodology originated with the Research Libraries Group (RLG) in the United States in 1979, and was adopted by the Association of Research Libraries (ARL) in the early 1980s, as a means of providing a map of library collections and collecting policies with individual libraries or among a group of libraries. It was also adopted in Canada where the National Library of Canada organised a number of regional Conspectus projects. In 1984 the British Library began a Conspectus survey within its own collections, and the results were published in 1986. Conspectus methodology was also applied among Australian research libraries. Scotland decided to use Conspectus in a group project in 1985, and this was the first group use of the methodology in Europe. It was then adopted in the Netherlands, based very much on the Scottish model of use. It has also been used to a more limited extent by some libraries in other parts of Europe (e.g. Sweden, Italy and Spain).

3.1.2 Aims

The original aim of the methodology was to try to find a mechanism that would enable librarians to 'measure' the strengths and weaknesses of their collections. There was a dual aim of identifying strengths and pinpointing weaknesses, so that the strengths and weaknesses of collections could be viewed not only within the framework of an individual library but among a group of libraries or across a region or a country. The primary purpose was to assist co-operation among libraries so that limited resources could be used more effectively to maintain services to researchers. Even then it was clear that the resources available to libraries to build collections and maintain services were under pressure, and likely to continue to be so for the foreseeable future. At its inception there was also a general view that if this could work to assist libraries to share resources for collection building and maintenance, then it might also be capable of use for sharing preservation responsibilities, and that there might also be other potential resource-sharing applications.

3.1.3 Methodology

The original Conspectus methodology provided a subject framework, issued in the form of worksheets, of 25 Divisions based on the Library of Congress classification system (e.g. Art and Architecture, Economics and Sociology, Medical and Health Sciences). These Subject Divisions were further sub-divided into about 250 categories (e.g. Architecture, Sculpture, Painting etc. within the Division of Art and Architecture); each category was then sub-divided into individual subject lines (e.g. Ancient, Mediaeval, Renaissance. Baroque etc. within the category of Architecture), against which the Conspectus level indicator was applied. There are around 7,000 subject lines in total.

The Conspectus level indicator is an alpha-numeric code used to indicate both the existing collection strength (ECS) of a subject, and the current collecting intensity (CCI), i.e. what a library already holds, and what it is trying to achieve in terms of collection-building.

The Level Indicator has six levels:

The Language Coverage codes are:

In addition to the Level Indicators, there are two free-text areas that can be used to supply additional information. Scope Notes are used at the Division or Category Level to give an overview of a library's collecting policies; Comments are used at the subject line level to supply a note on a particular collection.

3.1.4 The use of Conspectus in Scotland

The decision to adopt the Conspectus in Scotland programme took place within the then Working Group on Library Co-operation, now the Scottish Confederation of University and Research Libraries (SCURL) in 1985. The members involved were the then eight university libraries (Aberdeen, Dundee, Edinburgh, Glasgow, Heriot-Watt, St Andrews, Stirling and Strathclyde), the major public libraries in Edinburgh and Glasgow, and the National Library of Scotland.

The aim of adopting the Conspectus methodology in Scotland was influenced by the following factors: a desire to build up a co-ordinated 'national' resource in Scotland; the recognition that resources were under pressure and likely to remain so; a desire to provide the best possible services for our users from the available resources; a desire to establish the collective strengths and weaknesses of our respective collections; a desire to find an existing methodology to utilise if one existed. A primary recognition in Scotland was that the Conspectus methodology was not perfect, but it existed, and we would not have the resources ourselves to devise an alternative system.

A major decision in our application of Conspectus was the adoption of a fast-track approach to Conspectus. This had the aim of completing the survey within one year in order to give a general map of strong and weak collections and collecting intentions, rather than taking a more fundamental approach to the methodology. From our earliest examination of the methodology we had recognised that it was not a perfect system (as had also been recognised in North America). However we also recognised that it is unlikely that a totally satisfactory system for measuring collections could ever be devised. Since we wanted to have a fairly practical overall map of Scotland's strong and less strong research collections as a basis on which to carry out subsequent work, we came to the conclusion that we should utilise Conspectus in a pragmatic way on a short timetable and then refine and use the resulting information in the way that we wanted.

The programme began in October 1986 and was completed in September 1987. The outcomes from that programme were:

  1. A map of the collection strengths and weaknesses of Scotland's research libraries
  2. A clear understanding of the mission of each library in its collection-building
  3. An understanding of the active current collecting policies of each library in each subject field
  4. An overall picture of the weak areas of Scotland-wide collecting.

In terms of the resources used by the initial Conspectus assessment, the participating libraries (11 in all) took a total of 4,500 hours; and at the time the costs were estimated collectively by the participating libraries as approximately £35,000.

Having achieved these objectives from Stage 1, the next stage of co-operative work was:

  1. to accept that, given the financial climate, it was unlikely that weak areas or absence of coverage could be actively strengthened in any way by assigning more resources (except where a university library, for example, might build up specific subject areas if new teaching courses were introduced for some reason); and
  2. that having identified where there were strong collections, we should move to a position where we had examined these collaboratively and then tried to work towards a system of distributed responsibility for the whole of Scotland.

Stage 2 was the Co-operative Collecting Responsibilities (CCR) Programme, which was implemented in 1990. The objectives of this Programme were:

  1. To ensure that research-level collections in given subject fields were maintained in Scotland
  2. To engage in a voluntary programme of accepting co-operative collecting responsibilities which would identify in all subject fields the number of collections that were actively to be maintained
  3. To enable individual libraries to make decisions about their own collection policies in the light of knowledge about other libraries' intentions. The National Library might, for example, decide not to continue to collect at research level in a subject field if it became clear that there was already good coverage across the country (given that there were no additional factors to take into account) and might then build up another subject field for which there was a collective demand
  4. To find a way to keep this information up-to-date.

The Programme worked in the following way: all libraries with a Level 4 or 5 Current Collecting Intensity (CCI) were asked if they wished to accept a CCR in that particular subject. Where no Level 4 or 5 collections were recorded, a group of up to three Level 3 collections were asked if they wished to accept a CCR as a group. The theory behind this was that US experience had shown that three libraries, each with a Level 3 collection, tend to be so diverse in actual titles as to represent collectively a research-level provision. The principles underlying the implementation of the CCR Programme in Scotland were:

  1. Accepting a co-operative collecting responsibility (CCR) did not commit an individual library to any additional expenditure: it simply involved a commitment to continue to maintain a research level collection.
  2. The only additional responsibility was if a library found that it could not continue to maintain a research level collection for which a CCR had been accepted, the library must inform other libraries in the Programme.

The co-ordinating role in the implementation and monitoring of the CCR Programme has been taken by the National Library of Scotland.

Stage 3 of the Programme, following completion of the CCR Programme, was to mount the information about the existence and location of research level collections in Scottish research libraries online. The objectives of this Programme were:

  1. To provide researchers with information about where research level collections in specific subject fields could be found in Scottish research libraries
  2. To show which libraries had undertaken to maintain specific collections at research level
  3. To keep the information up-to-date to reflect any changes (expected to be minimal) in the collecting intentions of libraries in the programme
  4. To find a way to enable individual libraries to make any changes locally attaching a note of the date of the change, with overall co-ordination of the system being provided by NLS. The data has now been mounted by Strathclyde University Library as Research Collections Online (RCO). Some editing work has yet to be carried out and the intention is to try to find a means to enable this work to be implemented as soon as it can be arranged.

The general approach of the current level of work is:

  1. RCO provides researchers with information about where research level collections are located; and
  2. the online catalogues of individual identified libraries can then be searched for item level titles or subject searches.

Clearly, until retro-conversion has been completed in all the libraries involved, this will not work comprehensively, although that is the ultimate aim. The main objectives of this Programme now are:

  1. To keep the information up-to-date with any changes being made locally at the level of the individual library, with overall monitoring from NLS.
  2. To add research level collections maintained in the 'new' Scottish universities that had not been created at the point of the original CCR Programme.

A new stage of work is to identify a means of using the information gained through RCO as a means of sharing preservation responsibilities among Scottish research libraries (through the Shared Preservation Project that is currently being devised).

Another development of relevance is the eLib-funded CAIRNS Project. Consideration is being given to the future relationship between RCO and other relevant Scottish resource-sharing programmes, and the CAIRNS Project.

Ann Matheson, National Library of Scotland

3.1.5 CAIRNS, RCO, Conspectus, Collection description

CAIRNS [CAIRNS] aims to integrate around 25 Z39.50 compliant catalogues or information services of CAIRNS sites across Scotland into a 'functional and user-adaptive' test-bed service. This will offer search and retrieve capabilities across a 'clump' of services comprising all of the individual CAIRNS bibliographic databases, together with various services describing electronic resources.

The project will take advantage of SCURL's Web/Z39.50 Conspectus-based research Collections Online (RCO) service as the basis of a subject-based dynamic clumping service. This will provide users with dynamically generated subject-based 'sub-clumps' of selected CAIRNS services to search via Z39.50. The feasibility of extending the data to:

will also be considered, although in some cases, outwith the scope of the project. In addition, the recent availability of BUBL's 5:15 service based on the JISC Committee on Electronic Information (JCEI) subject categories may offer possibilities for examining possible mappings between Conspectus and the JCEI subjects.

The RCO service has only recently been made publicly available by SCURL [SCURL]. The service allows users to search SCURL's collection description database via the Web or Z39.50, although the initial structure of the database is more suited to Web access and the way that the data is held may have to be re-worked to take account of the requirements of utilisation within CAIRNS, particularly in respect of Z39.50 search and retrieve operations. The RCO service as presently implemented provides in-depth descriptions of collections held in the eight 'ancient' and 'modern' universities in Scotland, together with those in the two main public libraries and the National Library of Scotland.

Although Z39.50 access is possible, RCO is currently designed mainly for Web access. Users can

The result of a search or browse operation on the service is typically a record that looks like this:


Aberdeen University 
Level: 3

Glasgow University
Level: 3

National Library of Scotland, Edinburgh
Level: 3

[Note: Level 3 = instructional support; Level 4 = research; Level 5 = comprehensive]

Having retrieved such a record identifying (presumably) libraries holding collections of interest, the user may then click on the name of each library in turn in order to search the local catalogue. There is an assumption that geographical considerations may come into play at this stage and that in the above example a researcher in Aberdeen (for example) may elect only to search Aberdeen University and the National Library of Scotland in Edinburgh and not Glasgow University. Whatever may be the case in this respect, the user, having identified the library catalogues that are of interest to her, must then effectively exit from the RCO system and search the local catalogues individually. Thus, in the example above, she would connect to each of the catalogues in turn and repeat her subject or named item or class search in each, retrieving three distinct result sets in three distinct systems, using three quite different user interfaces. There would of course be the appearance of a unified system since most of the systems concerned are accessible, like RCO itself, via a Web browser. But there is no real integration, as is clear from the fact that the result sets are difficult, if not impossible, to combine together into one set.

The proposed CAIRNS dynamic clumping service would build on the present service and allow the user to conduct a subject-based collections search as above. However, retrieval in this case would take the form of a dynamically generated clump of appropriate Z39.50 servers that could then be searched simultaneously as a group using a broadcast search facility. The initial aim is to improve upon what is possible at the moment with RCO. Once the RCO search is completed, and the clump of servers retrieved, it will be possible to search all of them in what is effectively a single operation - so that users seeking to locate works of interest to them in 'collections' distributed across Scotland could, in effect, search a single virtual catalogue of the 'collection' and retrieve a single set of results (as opposed to two or three or five, which is the current position). The wider aims of the project include:

CAIRNS research in these areas is at an early stage but the following are likely areas of investigation and interest:

3.1.6 Retrieval within 'Collections'

There is currently no direct connection between an RCO record describing a collection and the items within an individual catalogue that 'comprise' that collection. The RCO record contains no information about the catalogue record and the catalogue record contains no information about its RCO subject categories or levels. Having identified a catalogue or set of catalogues containing a 'collection' of interest, the researcher might reasonably assume that her subsequent search will be carried out only within the 'collection' she is interested in, an assumption that might well affect her search strategy and results. However, this is not currently possible and, indeed, may never be possible. Whether or not it is important that it should be possible is still a matter of debate within CAIRNS, and it will be interesting discover whether user evaluation of the service clarifies the issues in this area. If it does, this may help identify desirable improvements in the way Conspectus terms relate to local records in the Scottish context, although such improvements are only likely to be implemented if they do not have a negative affect on the ease of maintenance of the Conspectus data.

3.1.7 Scalability

There are Conspectus-based projects in a number of other countries (see below). However, there is no reason to believe at the moment that Conspectus is likely to gain universal acceptance as an approach to collection description, even within research libraries, nor is it necessarily the case that it is universally applicable in any case. A major concern within CAIRNS (and, indeed, within eLib) is that mechanisms put in place as a result of the work of the project be scalable. This, in the context of RCO and Conspectus and dynamic clumping, includes not only technical scalability but also universal applicability and the likelihood of some reasonable level of universal acceptance. In respect of these considerations, questions for CAIRNS to consider include:

3.1.8 Explain

Nothing that has been undertaken so far in relation to SCURL's work with Conspectus and RCO has been undertaken in the context of any possible future relationship with the use of the Z39.50 Explain service. This has an important bearing on the design of the dynamic clumping service and will focus attention on a number of areas:

There are almost certainly other issues, but these are some of the main themes. There is little likelihood that CAIRNS will resolve all of the associated questions satisfactorily in its two-year lifespan. More probably, the best that can be hoped is that it will clarify one way forward in respect of collection descriptions and will identify the options that exist as regards their use and usefulness in dynamic clumping and distributed searching.

3.1.9 Other Conspectus-type activity

The countries that have taken the original Conspectus methodology and built upon it are:

The Netherlands
Conspectus values are added at item level in the PICA database (the union catalogue of The Netherlands). The Dutch have also implemented a co-ordinated resource-sharing programme, and in a sense have followed the Scottish approach fairly closely.
Australian research libraries have used Conspectus to assist them to develop a co-ordinated 'national' collection. They are also currently looking at its potential for sharing preservation responsibilities.
A pilot was carried out about two years ago, and there is now interest in developing this approach further. They are in touch with the Scottish approach.
A Conspectus exercise involving Welsh research libraries has been completed.

A number of other countries began to use Conspectus methodology and have not continued to pursue it for a variety of reasons:

A lot of resources were invested in the Canadian Conspectus programme but a fundamental approach was taken to its implementation and so the process took very much longer. It appears that the programme got waylaid and nothing further was done for some time. However in autumn 1997 there was some general discussion about taking up the idea again, as a result of the severe funding cuts experienced.
Considerable effort was expended in translating the worksheets into French (with the assistance of the Canadians). However, when the new BnF plans were projected, the priorities for French libraries were changed internally (by the Ministry of Education) to retroconversion of catalogues and digitisation.
A number of libraries and groups of libraries have applied Conspectus and some individual libraries have taken the results either for resource-sharing activities or for preservation. However, there were some practical issues regarding the 'ownership' of Conspectus between ARL and RLG. In addition, RLG priorities have moved recently very much in the direction of digitisation and digital projects.
British Library
The BL carried out a Conspectus-based review of its own collections in the 1980s, and it appointed a National Conspectus Officer to co-ordinate activities across England (and, indeed, the UK). The lack of interest from English libraries caused them to phase this post out about two years ago, and it appears that they are not keeping their own information up-to-date.

3.1.10 Results from Conspectus implementation

The main point about the Scottish use of Conspectus is that it really provided a foundation to begin the work of identifying research level collections. In itself it did nothing more than that. The important aspect of the Scottish programme has been the further stages to which we moved from this first level mapping made possible by the use of the Conspectus methodology. This has enabled us to:

  1. Agree a co-ordinated programme for the maintenance of research level collections across the country, and a means of sharing this responsibility among our research libraries.
  2. We have created subject specialist groups (e.g. Official Publications, Visual Arts, North American Studies) who have taken the primary Conspectus information and used it (along with other information) to implement more detailed resource sharing activities where these are possible; and to promote the collections in these subject fields more actively for the benefit of researchers.
  3. To form a possible basis for further resource sharing such as responsibility for preserving collections.
  4. We have had experience of collaborating through a programme that tested us quite thoroughly and gave us a good foundation for future collaboration.

We do not expect to maintain the base line Conspectus information any further since its purpose is now superseded. Our main interest now is in Research Collections Online, keeping it up-to-date and in developing other resource-sharing activities such as sharing preservation responsibilities.

3.1.11 Overall findings: Conspectus

  1. The original Conspectus methodology is very time-consuming and labour-intensive if it is approached in a fundamental way.
  2. It is our contention that you do not achieve a better quality of information by taking this approach, than a 'fast-track' approach.
  3. In our experience the useful aspects of the initial Conspectus methodology are the Level Indicators and the Language Indicators.
  4. In our overall use of the methodology and its aftermath, we have found that the most useful aspect is the second stage of the Co-operative Collecting Responsibilities (CCR), if the objective of a group of libraries is to identify where strong collections are to be found and maintained across a region or country.
  5. There are important advantages to research libraries, in the current funding climate, if they are able to show that they are actively collaborating to make the best use of the available resources for their researchers. The use of Conspectus, and even more so the CCR Programme, enabled us to do this in Scotland and this has been beneficial.

Dennis Nicholson, Strathclyde University