Sweden now has a comprehensive system for nationwide library statistics from all publicly funded libraries. This is how we did it.
The reason for the changes to all library statistics was primarily to allow us to reach the level of quality that Swedish official statistics are meant to maintain. A new statistics law was enacted in 2013 which raised the quality requirements and also made it mandatory to abide by the European Statistics Code of Practice.
The previous model for collecting library statistics was unable to meet all the requirements fully. Among the reasons for this was the fact that the data was collected in four different surveys, and many of the questions and definitions in these varied depending on which type of library was responding.
These surveys had evolved gradually over a number of years, and in the end library loans, for example, were different in all of them. The sampling frame was likewise ‘flexible’, and it was no longer evident why some libraries were sent the survey and others not. Some types of libraries were never approached at all.
Since Swedish libraries are increasingly consolidating into regional systems and integrating several library types in a single operation, our assessment was that it was no longer clear that different library types need different questions. Upper secondary school libraries, for example, are often equivalent in their activity to smaller research libraries.
Public and school libraries in many municipalities are part of the same library system. Compulsory level in the survey has been adapted to the level that school library systems are normally capable of delivering.
The same questions to all
We focused on what type of data users of the library statistics need, taking into account that there are many different types of users. Questions asked in the new statistics system are now based on the content of media more often than on which specific storage media type it uses.
New types of media are being developed all the time, and in order for the statistics to continue to be applicable for many years to come, questions now refer to “Film, radio and TV” defined as: “Media containing moving images or sound which is not purely a music recording”.
Children and young people
The focus has shifted from whether it is a phonograph cylinder or a DVD. We decided that it was more interesting to see what type of content the users wanted, rather than what type of technical formats there are.
However, we have not been able to resist “Books with written text” and “Audiobooks” as well as “DAISY talking books” either on a physical medium or digitally.
New questions have also bee added as a result of Sweden’s new Library Act from 2014. The act prioritises certain groups of users, and for that reason there are now special questions about children and young people, people with print disabilities, and access to media in foreign languages and national minority languages. The proposal presented by the Expert Group on Library Statistics also influenced the design of the survey.
Collaboration on the new survey
First we drew up a suggestion for a new survey and then we contacted organisations, authorities, interest groups, library employees, expert groups and individual library users for their opinions on the new survey. And did we get opinions! We held meetings, invited people to discussions, and received feedback via all channels.
There was a surprisingly large amount of interest in library statistics. The opinions that people contributed influenced the final design of the survey. Definitions were adapted to the international standards for library statistics (ISO 2789 and ISO 11620) where possible under Swedish conditions. The National Library has contributed to the development of those standards.
Discussions about the sampling frame
We limited the sampling frame to include all publicly funded libraries which are in some way accessible to the general public (in school libraries: the pupils) and which have staff allocated to them for at least 20 hours per week of library activities – this in order to achieve a fixed sampling frame.
There were several reasons for this, including the wording of our agency directive, the Library Act, and the consideration that library staff should have a reasonable opportunity to answer the questions. The new frame also makes it easier to monitor how staffing is being undercut at libraries in many parts of Sweden.
School libraries, which typically have very limited staff or none at all, were the main casualties of the sampling frame. Suddenly it looked as if some municipalities simply did not have a school library – and the fact that the National Library had set a minimum staffing limit for statistical purposes caused a public debate in Sweden about what reasonable staffing allocations are for school libraries.
Sweden’s Library Act and Education Act both stipulate that all pupils must have access to a school library. But can an unmanned bookshelf really be called a school library? We saw that our task was not to establish whether there are books in Swedish schools – we already know there are – , but to measure school library activity, which requires allocated library staff.
He who seeks finds
Sweden has never before had a complete register of all publicly funded libraries, but through extensive surveys we created one. For example, we asked close to 6,000 school units if they had a staffed school library, and found 820.
It was harder to find libraries than we could have imagined. In the end we came up with just over 2,200 libraries with unique street addresses, that fell within the sampling frame.
In order to improve the quality of the collected data, a customised web-based collection system was devised with validations and automatic summations. The system also has easily accessible definition texts by each question. The answers are routed to a database which then exports and generates activity reports with diagrams, key figures and tables directly on bibstat.libris.kb.se/report.
In Sweden it is a statutory requirement that all municipalities and county councils have a library plan, and one purpose of the activity reports generated by the system is that municipalities can use them when following up the library plans.
In order to facilitate things for respondents, and avoid double counts at the same time, all the libraries in one municipality can report jointly in web survey, provided they all have the same library system.
Each web survey displays a list of all libraries within the same municipality. Those reporting jointly can be selected in the list, thereby disabling their possibility of submitting answers via another survey.
Since many libraries chose various ways of reporting jointly, the number of surveys was reduced – and thus the overall information burden for Swedish library statistics as well. The data was collected by means of about 1,200 surveys in total.
Most things went well, but not all
We are surprised at how closely this year’s total figures matched those of previous years, despite the radical changes to the survey. Suppliers of Swedish library systems were not given enough time to alter the statistics extracts from the library systems before the new survey was sent out,
and many of those who provided data had
to make a considerable effort to obtain the
new type of figures.
During 2015 a further development of the library systems’ statistics clients will therefore be carried out.
There was a slight decline in the number of responding libraries, despite Swedish Official Statistics responses being mandatory. We also had a few technical mishaps along the way, but since the surveys went out to the most patient and informationcompetent professional group there is (yes, librarians), we received lots of replies all the same.
A very big thank you to the country’s statistics librarians, who managed to come up with what we were asking for despite everything being new and unfamiliar! We meet again next year…