Friday, 4 February 2011

Flawed Statistics

Some of you may be interested in Martin's letter(below) re the County's flawed statistical approach

Cllr Keith Mitchell
Leader of the Oxfordshire County Council
County Hall
New Road

Dear Cllr Mitchell

Re. Proposed Library Closures and concerns for manipulative use of statistics

Few sane people are in any doubt as to the need for state budget cuts at this time, and I think that most of us have shown willing to tighten our belts when it comes to reductions in the provision of many state services. Though a regular library user myself, I do not see the libraries as a sacred cow with this respect. However, I do believe that these services should be cut in a transparent and un-manipulative manner which shows respect to all tax-payers. When I attended a Parish Council meeting joined by three representatives from the County Council (Judith Heathcoat, John Jackson and Arash Fateman), back on 9th December 2010, I was disturbed by the strange survey methods used to justify the currently unshakable stance of Oxfordshire County Council – to completely close 20 libraries and leave 23 open.

My work in monitoring and evaluation regularly requires me to make statistical analysis of sociological data, and to do so I must establish a solid foundation of common-sense around which the maths and statistics are anchored. It is essential to establish what ‘units of measure’ best represent the object of the study. Yet this approach seemed wholly lacking in the Oxford County Council survey figures shared by John Jackson in that December meeting. I was astonished to learn that you had carried out a survey of library usage which took no account of the respective size of the library or threshold population that was being surveyed – hence Abingdon library, with a threshold population of around 40,000 people was compared like for like with Kennington Library with a threshold of only a tenth that size. In other words the ‘units of measure’ – which were ‘number of users per hour of library opening’ – were ill-founded and misrepresentative. Of course such a survey will conclude that Abingdon library receives more custom! What else would you expect to discover! Any survey worth its salt must take into account the threshold population or it is not worth the paper it is written on. The stats shared that night stated that Abingdon received an average custom of 130 users per hour of library opening – compared to 17 users per hour of library opening in Kennington. If one were to review the maths such that the equations actually held up to scrutiny, then the figures would be measured using the units ‘number of users per head of threshold population per hour of library opening’ – more of a mouthful, but far more representative of the true picture. If these units were applied to the most recent census figures (2001 census: Abingdon 36,626; Kennington 3,881) then the figures would adjust as follows:

Abingdon library: 3.55 1
Kennington library: 4.38 1

1 = “users per thousand of threshold population per hour of library opening”

By failing to establish a survey method that holds up to scrutiny, I proffer that the council is lying to its constituents by misrepresenting the truth. Whether this is a deliberate act of subterfuge, or simply an act of incompetence, only you can answer. But either way, if false statistics are being used at the heart of the argument for rural/suburban library closures, then it is essential to establish the truth of the matter.

I am a father of two boys under 4 in age who love the local library, and who would rather go to the local library than the swings (and this seemingly through their own choice not mine). I would much rather see our local library remain as a professionally maintained local service than see library functions centralised to Oxford and Abingdon – which I am not going to realistically ‘pop in to’ as I and my four-thousand fellow villagers currently can in Kennington. Therefore I advocate that the 25% cuts are made across the board – recognising that fixed costs will mean that a 25% cut will result in maybe a 30-40% loss of service. This is the only way to ensure that the maximum number of local children will develop a love of books at this critical age before school starts.

Please reconsider the methods you are using to define what makes up an essential library service, and apply your cuts equally across all libraries. We as a community recognise that there will be costs that need to be picked up and are willing to look into ways of addressing such shortfalls – but we can only do so on the foundation of a professionally run library at the heart of our community.

Thanks you for listening.

Yours sincerely,

Martin Hull


  1. It will be interesting to see how they answer these points!

  2. Brilliant letter. Very well-argued.
    They really ought to listen.