But all statistics are philosophical. Although the word can be used for a branch of pure mathematics, it is mostly used in an applied sense – figures which come with words attached
And then, in the enduring words of Professor Joad, it all depends on what you mean
I was taught economic statistics by the late RGD Allen, who emphasised the importance of never leaping in to analyse or draw conclusions from a set of figures before checking the definitions employed & any caveats attached due to the limitations of the method of compilation
If something is to be counted you need a careful definition of exactly what is meant, or qualifies to be counted in a particular category. This is especially true of statistics in the original sense of state figures, produced by official statisticians. Much effort goes in to this
There often has to be negotiation with a whole range of users to ensure that the figures are fit for purpose, will meet a wide range of needs. Sometimes multilateral negotiations are involved, for example in the European Union or for the even more majestic International Classifications of Disease or Trade
A major downside of all this is that the definitions are then rigidly applied & used. In written reports or press releases there is no room for elegant variation – the words should say what they mean and mean what they say. Any percentage or rate quoted should spell out both the numerator & the denominator being used
Nothing like what happens in normal life, where words mean what you want them to mean, & what you mean is obvious to your readers & listeners. Especially in political discourse
Words have a nasty habit of changing their meaning over time. One of the most difficult professional judgements for statisticians who collect figures over time is when to change the definition so that the statistics remain meaningful for society as it has now become. The price for this is a break in the time series & added problems in comparing the situation now with what it was then. They do not teach equations or formulas for this in university – you just have to learn by experience to use your professional skill & judgement
One of my favourite examples is lone parent. From at least the 1930s through the 1970s the definition adopted for the published census tabulations of households was one developed by social scientists for measuring the adequacy of housing provision
A lone parent household was one “Headed by a single/widowed/divorced parent living with his or her never married child(ren) of any age”
So a 90 year old widow living with a 70 year old spinster daughter counted as a lone parent in an age when “lone parent” was becoming synonymous with teenage girls who just wanted a council flat of their own
Politicians & others were quick to criticise another set of official statistics in the early 1980s. This was the time of the Thatcher recession when there were suddenly large numbers of (mainly young) people sleeping on the streets of central London in a way which was distressingly visible. Ridiculously enough, such people were deliberately excluded from the official homelessness statistics. In this case however the statisticians were able (gently) to point out that they were merely charged with monitoring the workings of the Housing (Homeless Persons) Act, passed by Parliament in – oooh – 1977. For a time the word ‘roofless’ acquired a certain currency
Statisticians face different challenges when, instead of purpose designed censuses & surveys, they rely on administrative sources for their raw material. There may be breaks in a series if new procedures lead to differences in the counting method or its reliability, or because of changes to the rules of entitlement
“Fiddling the figures” then becomes an easy taunt – ask anyone who remembers the debate over unemployment figures in the 1980s as various schemes were introduced to give people some kind of training or something else to keep them occupied & away from the dole queues. We may be in for the same kind of thing again as the government introduces initiatives such as internships for new graduates to meet the current emergency
Then of course there is the question of foreign workers. Anyone can recognise one of those when they see one; the problem is to find a definition which can be applied to meet a wide range of circumstances, & then to find an accessible source of the up-to-the minute information required.
People enjoy entering into this debate when it comes to questions about who is, or is not, a foreigner when it comes to playing football or captaining the England cricket team. Somehow, though, the nuances are supposed to be magicked away when it comes to the more workaday world
Link
Related posts
