As of 22 April 2009 this website is 'frozen' in time — see the current IFLA websites

This old website and all of its content will stay on as archive – http://archive.ifla.org

IFLANET home - International Federation of Library 
Associations and InstitutionsAnnual 

64th IFLA Conference Logo

   64th IFLA General Conference
   August 16 - August 21, 1998


Code Number: 052-134-E
Division Number: VI.
Professional Group: Statistics
Joint Meeting with: -
Meeting Number: 134.
Simultaneous Interpretation:   Yes

UK special library statistics : the challenge of collecting and analysing data from libraries in the workplace

David Spiller
Library and Information Statistics Unit (LISU),
Loughborough University
Loughborough, UK


The paper describes some of the particular problems of surveying special libraries, based upon three recent pieces of work involving the Library and Information Statistics Unit. Difficulties of defining and identifying special libraries are described, and there is discussion of sources of information, and the classification of special library sectors. The reasons for poor returns of postal questionnaires are analysed. There is a section on the advantages and disadvantages of different methodologies, touching on sampling, and ways of eliciting and presenting quantitative data.



In most countries, special libraries are the part of the library and information world about which least is known. Even the label 'special library', though generally understood, is a nondescript term, defining in terms of what it does not represent (public, academic or school libraries) rather than what it does. A better description may be 'libraries in the workplace': services in commercial, governmental and quasi-governmental bodies, providing information to further the goals of their organisations.

The lack of focus is inherent, given the wide divergence in size, subject matter and type of the different special libraries and information centres - not to mention a natural unwillingness of commercial organisations to make information about their activities publicly available. But confidentiality can be preserved, and special librarians need not opt out of the powerful national and international trends towards benchmarking performance, in order to establish good practice and improve their services. In fact, given the economic and social importance of the institutions, and the importance of information to their activities, an overview of the quality of their information services may be urgently needed. If reliable, sector-wide information can be obtained from, and made available to, the managers of special libraries - and their managers - it will also prove invaluable to those who supply these sectors with services: materials and electronic resources, equipment, training and education.
In this paper I should like to suggest why the sector is difficult to come to grips with, and to describe work done in the UK over the past decade and the methodologies which have proved most effective.

Recent UK work

In the UK, three pieces of statistical work on special libraries have been carried out in the past five years:

  1. UK special library statistics (1), by Pamela Berridge and John Sumsion, was a survey of the statistics on special libraries that existed at the time of publication (1994). It was funded by the British Library Research and Innovation Centre. The exercise was crucial in paving the way for subsequent surveys, and it may well be that this kind of mapping approach is an ideal start for a country where the subject of special libraries is being addressed for the first time.

  2. TFPL survey of UK special library statistics (2), by Claire Creaser and David Spiller, took advantage of information collected in 1994 and 1996 by the prominent UK information consultants, TFPL.

  3. Libraries in the workplace (3), also funded by the British Library Research and Innovation Centre, is a project in progress: a random survey of ten special library sectors, which builds upon the findings of the TFPL survey - particularly in the areas of performance and the use of electronic media.

A summary of the 1994 estimates is given in Table 1. Although very broad brush this gave at the time a new and valuable perspective on the relative size and activity of types of special library in the UK.

  	Table 1  UK Special Libraries - Sector estimates - 1994
			Number of 	Books 		Current 	Online  	CD ROM 
			units		acquired	periodicals	databases	titles 
									accessed	held
Pharmaceuticals		166		38,000		35,000		1,780		  600
Chemicals, Plastics	166		34,000		31,000		1,500		  750
Energy, Metals, Mining	130		34,000		43,000		1,180		  660
Manuf'g: machinery	239		44,000		33,000		1,310		1,460
Manuf'g: non-machinery	 75		22,000		15,000		  520		  240
Banking & Finance	168		24,000		28,000		1,500		  600
Legal			114		33,000		22,000		  720		  330
Consul'ts/Inf'n 	 51		 8,000		 6,000		  360		  570
Media			 60		11,000		 3,000		  260		  960
  TOTAL		      1,169	       248,000	       216,000		9,130	        6,170
Other libraries:					
Government 		590	       442,000	       220,000		  930		1,450
  & Agencies
Nat Health Serv		515	       308,000		56,000	        1,080		1,410
Professional 		200		90,000		58,000		  620		  840
Local government	118		41,000		17,000		  520		  340
Voluntary		420	       124,000		60,000		  760		  920
Museum			 60		90,000		42,000		  110		  500
  TOTAL		      1,903	     1,095,000	       453,000		4,020	        5,460
GRAND TOTAL	      3,072	     1,343,000	       669,000	       13,150	       11,630


Anyone undertaking a survey of special libraries will encounter serious problems of identification and definition. It is first necessary to decide which types of special library are to be included. The term 'libraries in the workplace', suggested above, automatically excludes certain categories - such as private libraries and libraries in clubs - and itself needs breaking down into sectors, so that like services can be compared with like. The classification devised by Pamela Berridge and John Sumsion, and used for the first two LISU publications mentioned above, has proved a workable system. The categories are:

For allocating industrial and commercial libraries to sectors, the list of codes in the 1992 Department of Trade and Industry 'Standard Industrial Classification' was used - an approach which eliminated much of the potential cross-classification related to subject. Even so, certain institutions obstinately resist easy allocation to one sector, for various reasons: the borderline between 'commercial' and 'non-commercial' becomes increasingly indistinct, as organisations which were 100% government-funded are obliged to diversify their sources of funding; in areas such as law and medicine, academic libraries are often hard to separate from special libraries, and listings often bundle them together - as, for instance, in the valuable directory produced by the British and Irish Association of Law Libraries (4); research libraries are particularly difficult to classify, and might almost warrant a separate category. One could go on at length about problems of this kind.

Once sectors are defined, it is necessary to identify all the libraries in organisations within them. This is a difficult step. Special libraries are rarely as well mapped as their academic and public counterparts. Those directories that do exist are rarely comprehensive. Directories of the organisations in these sectors tend to be more complete, but rarely give details about their libraries. Amongst the sources that have been found useful for the projects mentioned above are:

For some of these categories, the commercial nature of the parent company, and/or the confidential nature of the information held, means that access to information needs to be negotiated - sometimes with payment of a fee.

A recent paper by German colleagues (5) lamented that it was difficult to achieve returns of more than 35% in surveys of German special libraries. That experience is echoed in UK surveys. One reason for it is the difficulty of getting postal questionnaires to the right department of the organisation. The different terminologies used (library, information unit, resource centre, research, documentation centre, etc) do not help here. Telephone chasing of postal questionnaires is essential, and often establishes - if the right person can be tracked down - that the original questionnaire never reached its mark.

Of course, a questionnaire about library/information services is unlikely to be completed if the recipient organisation does not have a library. This can only be determined - if at all - by telephone chasing. In the Libraries in the workplace random survey, 14% of the 897 organisations contacted said they had no library, and the figure rose to 17% if the category included libraries which were too small to be surveyed. The full picture is shown in Table 2. Twenty-six of the 98 voluntary sector organisations contacted had no library, whilst amongst food manufacturing organisations there were 58 out of 97 - a response which suggests the need for a completely different survey about the nature of information gathering in manufacturing industries.

Table 2 Response rate

Code	    Sector	   Forms sent 	   Response rate	"No library" 
				out	        - %	          responses
1	Govt Dept		29		58.6		     0
2	Govt Non-Dept		96		45.8		     2
3	Voluntary		98		26.5		    26
4	Trade Assoc'ns		101		49.5		     4
5	Law			90		25.6		     8
6	Comm & Fin'l		99		24.2		    11
7	Energy			90		36.7		     7
8	Pharmac'l		97		33.0		     3
9	Man & Info Cons		100		33.0		     9
10	Food Manuf're		97		3.1		    58
Total				897		31.8		   128

Technical factors aside, there are other probable reasons for poor returns. In many special libraries there are only one or two hard-pressed members of staff, for whom the receipt of a questionnaire is not a cause for rejoicing. Libraries in commercial organisations are understandably reluctant to provide confidential information which might be of use to their competitors. (For this reason surveys should best come from outside the sector, from organisations with a reputation for not revealing sources.) Also, there is no strong tradition of benchmarking amongst special libraries - despite its prevalence in commercial organisations themselves. Perhaps, too, special librarians feel that the added value they give, by packaging or interpreting information for their users, cannot easily be addressed by a questionnaire - and on this point they are probably right.


The different approaches taken in the TFPL survey (2) and in Libraries in the workplace (3) are interesting and instructive. TFPL hit upon the notion of including some statistics questions within their professional/biographical questionnaire distributed to all those mailed for their publication Who's who in the UK information world (6). These were then developed in liaison with LISU. The impressive number of responses received to the statistical questionnaire - from individuals in 1,076 organisations in 1994 and in 772 in 1996 - was doubtless a function of this two-pronged approach, as well as of the brevity of a questionnaire squeezed onto a single page. With this approach extra work is required to avoid duplicating data from more than one individual respondent on the same LIS unit.

The current Libraries in the workplace survey (RR) is restricted to ten specified special library sectors, and respondents were selected using a simple random sample, drawn from a variety of sources. The questionnaire was longer (four pages) and the initial response much lower: 150 questionnaires, or 15% of the sample - a figure which increased to 31.8% following telephone chasing (not including those organisations which reported that they had no library).

Despite these differences of approach, the results - in areas where findings were available from both surveys - were largely comparable. Where they were not, the contradictions were mainly explicable by the differences of approach. For instance, the self-selection approach of the TFPL survey (addressed to individual professionals) naturally led to responses being received from organisations where a well developed library was known to exist. The random sample of organisations in Libraries in the workplace elicited information about organisations where there was no library, or where the library/information facility was very small - and many of the findings showed correspondingly lower averages.

Another difference was the form in which quantitative information (about budgets, stock, acquisitions, etc.) was requested. The TFPL survey used a bracket approach, asking respondents to tick one of five or six boxes representing different ranges of figures. This was probably a factor in eliciting a high response - particularly in sensitive areas such as budgets. But in devising the questionnaire, it is easy to guesstimate the prompted ranges unhelpfully, both at lower levels and higher (when a prompted option of '5,000+' can conceal figures of 5,000 or 50,000). Libraries in the workplace requested specific figures - more accurate, if they are forthcoming.

However it is gathered, quantitative information needs careful treatment at the analysis stage. The variations in responses tend to be much greater than they are in other library/information sectors, and almost every sector was characterised by a small number of very large libraries at the top end of the scale - and a consequent tendency for means to be distorted. For this reason, medians are a better indicator of the average in most instances.

Because surveys of special libraries are rare, the missing factor has been information about trends. The TFPL data, which relates to both 1994 and 1996, gives the start of a time series, and the Libraries in the workplace survey will establish a small panel of special libraries willing to provide information every year - so that some trend analysis should be possible within a few years. In the meantime, some indication of trends was obtained by questions in both surveys which asked whether budgets, staffing, acquisitions, etc. were 'increasing or decreasing'.

Some findings

From the Libraries in the workplace survey, I should like to comment on a handful of findings, drawn from the large amount of information that was collected. I hope that these may elucidate some of the points made above, and demonstrate one or two particular problems in surveying special libraries - and above all be of interest to special librarians.

First, the users of special libraries are both difficult to quantify and more diverse than might be apparent. One wants to know how far library/information services are of use to the staff of the organisation (on one or more sites), but also what use is made by a variety of external users: external members (e.g. 92% of professional associations served these); external clients (82% of government departments and 75% of financial libraries served these groups); and members of the public (69% of voluntary libraries served these). Information about user groups is presented in another form in Figure:

Fig 1 Clientele

The funding bodies of special libraries are also interested in the actual numbers of regular users, in relation to the numbers targeted by the library/information services. In Libraries in the workplace, the median numbers of users targeted varied from 13,000 for professional associations to 300 for voluntary organisations. The percentage of regular library users to targeted users varied between 46% in government and 24% in professional associations.

Expenditure per regular user
User data become even more interesting when linked with other information. Figure 2 depicts average expenditure per regular user. There are clear differences between the sectors, in overall levels of expenditure and in the relative levels on the four expenditure components listed. For instance, pharmaceutical libraries are the most intensive spenders, with serials by far the largest component. Financial libraries are the next highest spenders, though in their case most expenditure goes on on-line sources. The voluntary sector is the only one where the largest expenditure element is books.

Fig 2 Average expenditure per regular user


Methods of database access
We turn now to electronic information sources, which are usually assumed to affect special libraries more, and more quickly, than other types of library service. Figure 3 shows, by sector, libraries' reports on where use of electronic media was 'increasing' or 'decreasing' most rapidly.

Fig 3 Changing methods of database access


There were some interesting features here. Internet sources were frequently mentioned as 'increasing in use', as might be expected. The use of CD-ROMs was also reported as increasing in many sectors, but in both pharmaceutical and management consultant libraries it was decreasing more than increasing. Perhaps the most interesting feature was on-line services. One might have expected on-line services to have been the main casualty of increased Internet use, and in some sectors this was the case; but amongst management consultant, financial and voluntary libraries its use was increasing more than decreasing. The message seems to be that there are no straightforward patterns for the use of information materials, and that each medium finds its own level in specific situations.

When this approach is extended to the absolute numbers of searches we also find considerable diversity between sectors. CD-ROM is generally the most popular route for searching, except in the pharmaceutical industry, where on-line hosts are more frequently used. The energy sector makes greater use of the Internet than the other sectors. Management consultants make the most use of searching techniques in general.


I hope that this rapid summary of the three projects - their approaches, problems and methodologies - casts some light upon the particular problems of surveying special libraries. I believe the findings described here, and others for which there is not space to print, indicate the value of surveying in this area, and the need for further work. More detail can be gleaned from the publications themselves: the first two available currently from LISU (1)(2), and the third to be published shortly (3). The LISU web page, for latest information,
is http://info.lboro.ac.uk/departments/dils/lisu/lisuhp.html


  1. Berridge, Pamela J & Sumsion, J. UK special library statistics, ISBN 0 948848 64 2, Loughborough: LISU, 1994.

  2. Creaser, Claire & Spiller, D J. TFPL survey of special library statistics, ISBN 0 948848 93 6, Loughborough: LISU, 1997.

  3. Libraries in the workplace. LISU project in progress, publication Summer 1998.

  4. British and Irish Association of Law Libraries. Directory of British and Irish law libraries, 5th ed, Hebden Bridge, Legal Information Resources Ltd for BIALL, 1995.

  5. Deutsche Bibliotheksstatistik. Janke, E. "Revising the German special library statistics: a quantity -to-quality initiative, Discussion paper for the European workshop for suppliers and users of library statistics". In: Telematics for Libraries - From Quantity to Quality: Collection, Analysis and Use of Statistics for Libraries, Workshop Luxembourg, 9/10 December 1997. European Commission DG XIII - E/4.

  6. Nordin, Jorund B, ed. Who's who in the UK information world, 6th ed. 1997, London: TFPL, 1997.