Head of Stats warns politicians: 'Hands off our numbers!' UK Party leaders were this week on the receiving end of a sharp rap to the knuckles, as a letter from the Head of the UK Statistics Authority, the aptly named Sir Michael Scholar, warned them to keep their paws off official stats for the duration of the election campaign. The letter is a masterpiece of understatement, …
For an educated man he misses the point somewhat. The problem with politicians and statistics goes back further than Disraeli.
For every politician that quotes statistics to support his/her argument there will be an opposing politician who quotes a conflicting set of statistics to support the opposing argument. And both sets of statistics may well be true, for certain values of truth. And therein lies the problem. As Disraeli pointed out you can make statistics say just about anything you want and if you do it right you're not really lying, you're just not necessarilly telling the whole truth.
Political interference is a fact of life in his job, simply because the government will decide what is counted and how. It's always been thus.
As an example our current government measure unemployment in a completely different way to the previous incumbents (and to be fair the tories changed the way it was measured more than once). So the opposition could, if they had any sense, have spent the last thirteen years quoting the figures measured their way. Most of the time the tory way of measuring would have given a higher figure than the nulabour way. Now both those figures would have been right, they would just be measuring different things and giving them the same headline name. Every time people read the unemployment figure in the news they don't want to hear how it's measured, only statisticians and politicians really care. So people don't nessarilly know that people being paid unemployment benefit while on part time courses or in part time (even voluntary) work are not counted as "unemployed and claiming benefit". They are not generally aware that people who are not entitled to unemployment benefit may not be included in the headline unemployment figure, whether they are working or not.
The very fact that his particular office exists is down to political interference. It matters not whether we are talking about a national statistical body or some bloke measuring KPIs in the audit department of a company that make plumbing fittings all real world statistics are subject to political (not necessarilly party political) interference.
If he can't accept that he's probably in the wrong job.
Logic vs Statistics vs Politics.
Logic: At least one bear has shat in a wood.
Statistics: Surveying the habits of bears (as opposed to the incidence of ursine defecation in woods in general), at the 95% confidence level the presence of bears correlates with the presence of bear shit in bear-occupied woodland.
Politics:Unlicensed bears are fouling sites of outstanding natural beauty.
Party Politics: The (insert name) party is callously, wilfully, and cynically ignoring/exaggerating the issue of bear faeces and environmental degradation and the consequent effects on public health/education/safety/radicalisation of soft toys which we, the (insert name of a. n. other) party cannot condone.
Election Politics: The Bears are Coming!
Don't worry yourself about it Sir Michael. The electorate have long had the perfect answer to your problem. We don't believe statistics - period. Never have, never will. So when politicians bring out the latest spun figures to support whatever BS they're pushing this week, we simply don't believe them. It goes with the territory - their lips are moving, ergo they're lying...
Whether it's saving the taxpayer millions in benefit fraud (by including every single withdrawn claim) to NHS bed counts (including the chairs in the lobby) the reaction of most intelligent voters to statistics is - "Yeah... right! Says who?"
> the reaction of most intelligent voters to statistics is - "Yeah... right! Says who?"
Your statistic is misleading and is only true for a given value of "most intelligent voters" and therefore could be abused by unscrupulous politicians. Presumably your statistical sample does not include readers of the Daily Mail of the Sun but how can we be sure? You need to define how you measure your sample against the words "most", "intelligent" and "voters".
Let's start with the word "voters": does "voters" mean residents of the UK who are legally entitled to vote in the UK? Or do they also need to be a UK citizen? This is an important distinction because UK residents who are citizens of another EU state are entitled to vote in local elections but not general elections. And do you include UK citizens who live or are stationed abroad?
Now for the word "intelligent": how do you measure that? Do you mean people who have an IQ above 100?
And finally for the word "most": do you mean more than 50% of your sample?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
