Playing the numbers game in policy
The statistical illiteracy of the population – including policy-makers – was the subject of a discussion at the British Library on 13 April, as part of the Census and Society programme of events.
The journalist David Walker began the discussion with some observations and questions on the political role of the census, and what it says about British attitudes to the state. What does the state need to know? Does the state knowing more translate into better social and economic policy? Does it solve problems, or just present new policy challenges? He also asked whether a ‘Big Society’ needs a national census, or whether it shouldn’t, in fact, be ‘self-knowing’? And he asked if there is a correlation between data collection and big government.
Richard Alldritt of the UK Statistics Authority made the point that statistics are not facts. Statistics is the science of incomplete data – essentially estimates based on the best available information at the time. The challenge for statisticians is explaining the relevance, context and limitations of official statistics.
The second point he made was that it is essential to be able to trust the statistical service that produces the figures – official statistics should be produced using the right methods, and should be well explained. Trust, here, is not indivisible – it varies according to context.
He concluded that a sensible response to this is informed scepticism, rather than cynicism. Statistics inform and fuel public debate, but they don’t lead it. It isn’t a scientific discipline: each ‘side’ of an argument selects the statistics that support their own particular perspective or cause. And the UK Statistics Authority takes a dim view of the way in which some politicians use statistics that have no published source.
Yet it’s wrong to be too cynical. Richard Alldritt argued that if you believe nothing you’re told, your ability to make informed judgements is limited. The correct response is to question everything: who says so, how do they know, and what information has been left out.
The director of Straight Statistics, Nigel Hawkes, called for the honest presentation of statistical data by politicians and the media. He stated that good statistics underpin good governance – yet the most interesting statistical misdemeanours are often committed by government departments and non-departmental public bodies. He added that the charitable sector is particularly prone to statistical misrepresentation, in order to keep specific issues in the public eye.
How far should statisticians comment on public policy?
What is the solution to statistical illiteracy in the media and the wider population?
Is the national census the best way of collecting the data that helps social scientists understand society? Is there an acceptable alternative? Has the census passed its sell-by date?