- Picture: Petter von Krogh, Jannicke Røgler and Stig Elvis Furset from Buskerud County Library.
I’ll concentrate on two processes I’m personally involved in:
- improving the way library statistics are collected, processed and presented by library authorities
- systematic training in, and effective use of, statistics by the professional community of librarians
Change the system
- Improving KOSTRA (how Samstat confronted Statistics Norway). KOSTRA (below)
- Working with indicators (walk while they talk). Twenty indicators based on KOSTRA
- Access to data (Plinius Data). Explore and process.
- Shifting the field of discourse (from Pierre Bourdieu to Ray Lyons). Professional statistics (below)
- Statistics for Advocacy. Growing interest in Norway.
- Count the Traffic (everybody can be a researcher). CTT Home
- Statistics in Practice (SIP) (training takes time). Statistics Day 2011
Statistics are usually collected from librarians by library authorities.
Statistical systems tend to be designed by administrators. The balance of power favors the authorities. Most librarians are only moderately numerate. The libraries that collect and deliver annaul statistics every February, with much cussing and swearing, remind me of peasants paying tithe to the nearest church. They seldom see their statistics again, but they hope their annual gift of numbers will contribute to the general good.
The Church, hovever, is not very numerate either. It gathers the grain. It mills some of it. But it produces very little bread. Most of the corn is simply stored – unmilled, unbaked, unwanted. Administrators are neither millers nor bakers. Nor do the peasants care. They avoid the dark bread that statisticians want to make. Too much chewing. Let us eat cake!
The priests pretend they need the statistics. The peasants pretend to believe the priests.
- Statistics are good!
- Indicators are great!!
- Pie charts are glorious!!!
In the field of practice, it is business as usual. Numbers live in the sky. Librarians work on the ground.
Now, the balance totters. Officials and politicians start looking at the data. They ask unusual questions. How do you compare with KOSTRA group 11? They even make decisions based on statistics. We call this New Public Management.
The same thing happened many hundred years ago, when the first modern kings and their officials invaded the established feudal world. The rules of the game are shifting. The Norwegians use KOSTRA (ca. 2000-). The Normans used the Domesday Book (1086)
- PL 75/10: Digital and multicultural practices. Reflections on development work.
Many of the posts on my blog Plinius (English version) deal with library statistics – see Plinius on statistics. Quite a number deal with Norwegian library statistics (Appendix).
Norway is a country with a strong municipal sector. Public libraries are municipal services. With a population of five million, and more than four hundred muncipal units, most muncipalities are small. Many of them struggle economically and a growing number of public libraries is threatened with spending cuts.
Proposals are often based on comparative data on library costs and services available through KOSTRA, a massive nation-wide system of municipal statististics developed by Statistics Norway over the last decade. But the KOSTRA indicators for the public library sector suffer from several serious mistakes that render simple comparisons invalid: (1) some municipalities include, while others exclude, rent and infrastructure costs; (2) some libraries consist of many branches, while others have just a single central library; (3) some libraries are co-located with schools.
These technical issues are seldom taken into account, mostly because neither librarians nor municipal officers are aware of them. When people take the KOSTRA indicators at face value, the factual basis for decisions is faulty. The statistical measuring instrument itself is flawed. Since 2011 the pressure from “KOSTRA arguments” has been on the increase, and the interest in better methods of comparison is growing. The author has recently been invited by several county libraries to train librarians in the correct use of KOSTRA and other statistical data.
Source: Correcting the KOSTRA statistics
Most people accept that it takes training and talent to sing at a professional level.
In our cultures we are exposed to rock and pop and jazz and folk songs from an early age. We learn to recognize the difference between amateurs and professionals.
Everybody can play with numbers. Sort of. Real statistics is much harder. But very few people consider that it takes training and talent to do statistics at a professional level. They are unable to “hear the difference” between advanced beginners and real experts. …
When I say so in public, silence descends. I suspect Ray Lyons, my US colleague, has met similar reactions. But the time has come to speak out, as Ray has done – in Navigating with fragmentary information:
I have implied this in other entries in this blog, but I might as well say it outright: The library and information science profession needs to come to terms with the issue of standards for (i.e., rules of) evidence for performance, statistical, and advocacy research data. There, now I’ve said it …
At school and university students may get a cursory introduction to statistical methods. Our library students in Oslo devote thirty hours of their second year curriculum to research methods. This is intended to cover qualitative as well quantitative as well as methods. The course is fair enough as a brief orientation. But as practical training you might as well take a thirty hour course on modern European languages that covers French as well as German.
Professional musicians are recognized as such. Professional statisticians must shout to be heard. The things we have to say are generally unpopular. We attack the sloppy use of numbers, indicators, graphics and tables. We want clearer concepts, better data and stronger arguments. We ask for intellectual discipline and quantitative rigor.
Norwegian library statistics at Plinius
- P 8/12: News from the North
- PL 7/12: Electronic subscriptions
- PL 6/12: Correcting the KOSTRA statistics
- PL 3/12: Picture the problem!
- PL 35/11: Explore and process. About Plinius Data.
- PL 24/11: Size matters. Norwegian public libraries 2010
- PL 11/11: Immigrants in a digital Norway. Substantial study of digital skills
- PL 3/11: Whichbook and Ønskebok.no. Usage statistics in Norway and Great Britain
- PL 75/10: Digital and multicultural practices
- PL 60/10: Digital divide in the classroom. Solid Norwegian study.
- PL 55/10: An image of productivity. Using scatter diagrams to study public libraries
- PL 51/10: Student behavior in academic libraries. Systematic observation using CTT
- PL 50/10: Count The Traffic 2010. Revised presentation of the CTT method – with new data from public libraries.
- PL 47/10: Children who “live” at the library. Average visit lasts nearly an hour …
- PL 30/10: Lecture notes on CTT. A brief introduction to traffic counts in libraries
- PL 29/10: New fact sheets. Traffic data – and a bit of context – from individual libraries
- PL 16/10: Learn to count – by video. A fifteen minute introduction to Count The Traffic
- PL 14/10: Indicators and advocacy. Links to lecture notes in English and Polish
- PL 12/10: Fact sheets for libraries. Quick looks.
- PL 11/10: A simple way to knowledge. An article on the CTT study at Oslo University College – by three Erasmus students – has just been published by the Norwegian library journal Bok og bibliotek.
- PL 10/10: Size and productivity. Markku Laitinen comments on a study presented at Stellenbosch in 2007
- PL 6/10: Indicators based on KOSTRA. Proposal for a set of six external (advocacy) and twenty internal (management) indicators.
- PL 5/10: Indicator politics. IFLA-paper proposed for the Statistics and Evaluation Section 2010.
- PS 2/09. Private lives and public libraries. Proposed paper on traffic counting for Northumbria 8. See Florence 2009 for full paper
- PS 1/09. How to win friends. Proposed paper on statistical indicators for Northumbria 8. See Florence 2009 for full paper
- PL 52/09. Beyond the bubble. Combine library statistics with other types of statistical data
- PL 46/09. New indicators for public libraries. Two new papers on library statistics.
- PL 32/09. Computers everywhere. Mobile PCs on campus
- PL 1/09. Easy access to statistics. Announcing Plinius Data Bank.
- PL 49/08. Putting statistics to work 2. How to reason more with statistical data.
- PL 48/08. Putting statistics to work 1. How to reason with statistical data.
- PL 44/08. Visits in Denmark and Norway. Decreasing number of public library visits per capita since 2005.
- PL 42/08. Lending of non-fiction. New data on Norwegian titles.
- PL 41/08. Steady decline in book lending. The trend is down, down, down.
- PL 19/08. Physical and virtual traffic. My two papers for Quebec both deal with traffic..
- PL 7/08. Traffic papers for IFLA 2008. Libraries need visitors – and visitor statistics.
- PL 5/08. The high cost of quality. Proposal for IFLA satellite paper on technical and political issues in library indicator development.
- PL 4/08. Observing the traffic. Proposal for IFLA paper on physial visitor statistics.
- PL 3/08. How much is much?. Proposal for IFLA paper on national library web visitor statistics.
- PL 27/07. Ups and downs in lending. Monthly loans from Norwegian public libraries.
- PL 26/07. Close encounters of the library kind. What happens when people live far from libraries?
- PL 25/07. More fiction in smaller libraries. And more non-fiction in large ones.
- PL 12/07. Productivity and size in public libraries. Summary of paper for 7th Northumbria conference