John is a Conservative activist in Bosworth. He has had a career as a mathematician and manager in electrical, gas, and consulting and network communications and is a member of the Operational Research Society.
> Policy summary
Appoint Statistical Ombudsman.
> Policy explanation
The electorate are regularly subject to spin, where the powers that be select facts that support their message and totally ignore all the facts and analysis that would prove the spin wrong. It is not that the facts are untrue; it is the message that misleads. One well known area is road safety, where there has been an obsession with reducing speed limits to below the average observed and use of cameras to raise revenue. An established trend over several decades saw fatalities reducing by 3 percent or so each year – this appears to have ground to a halt under New Labour.
Remember when the General Medical Council struck off paediatrician Professor Sir Roy Meadow after his "misleading" evidence in the Sally Clark case. He was an eminent expert in his own field but was asked to give statistical/DNA opinion. The prosecution & judge should have realised he was not an expert witness regarding probabilities so should have asked for an expert in that field.
This policy would provide an avenue for any citizen to request a professional analysis to be carried out, with the Statistical Ombudsman able to demand the necessary data from the ministries, police and local authorities to get an un-biased analysis done.
> Political risks and opportunities
The Conservative Party has emphasised honesty; this policy would clearly show that we will put it into practice. The policy would encourage un-biased analysis before a bill is drafted. The downside is that many unfounded beliefs will be shown to be unsupportable by thinking people – didn’t everyone think that the world was flat etc.
> Questions for ConservativeHome readers
You may want to extend the range of the policy idea to other scientific and arts areas. The “Speed kills” statement has the merit of simplicity but little else. Accidents happen when observation, interpretation, judgement or control are lacking. Can anyone believe that terrorism has not increased as a result of an illegal, and illogical, war against Iraq?
> Costs
The cost of the policy will be a minor fraction of the savings achieved by not embarking on ill-founded actions.
Any initiative designed to bring honesty and transparency back into political debate after 10 years of spin must be welcomed.
When there is a TV programme on politics, it would be most refreshing to have a factual summary of the issue presented by a totally independent and non-political body before the politicians get on with trying to prove black is white.
Similarly in election campaigns, political parties should be barred from using their own stats.
The Treasury's Pre-Budget Report (now thoughfully available in doctors' surgeries) gives a vastly different reading of the economy overall than is provided by, say, the Telegraph.
Posted by: David Belchamber | January 09, 2007 at 10:03 AM
YES - I was wavering as I read through, but the last two lines put it all into perspective. This ties is nicely with an earlier 100 Policy on transparency in government spending.
Posted by: Mark Wadsworth | January 09, 2007 at 10:42 AM
I think you will find there is already such a thing. Its called the statistics commission www.statscom.org.uk which will investigate and report on specific complaints from the public as well as monitoring the ONS and Govt depts to ensure they play fair....it was set up by G Brown Esq.....another well researched policy from conhome then.....
Posted by: Wellness | January 09, 2007 at 10:45 AM
Yes this is important -and not simply to catch out spinning politicians.
Frankly, many people are innumerate, and the media like good (but statistically incoherent) stories.
Take this morning's example: the BBC is reporting that "adding milk to a cup of tea can destroy its ability to protect against heart disease, according to research"
The research in question involved studies on only 16 individuals. A sample size this small should not lead to conclusions such as this, but taxpayer-funded scientists and a compliant media frequently run such stories.
Perhaps this ombudsman could take a role in correcting such nonsense as well.
Posted by: Adrian Owens | January 09, 2007 at 11:16 AM
What need for an ombudsman when the underlying issue is general innumeracy? For an innumerate member of the public the pronouncement of the statistical ombudsman would be of no more weight or intelligibility than the spun figure from the body being investigated. For those who are numerate and already able to be sceptical about the figures provided to us there would be no value in being confirmed in our existing view by another public sector person.
Do we really need a Numbers Tsar? Can't we go and hug a Chi Squared Test without one?
Posted by: Angelo Basu | January 09, 2007 at 12:02 PM
In principle,yes.
In practice:No.
While legislation allowing access to unbiased statistics seems favourable; in practice, it would be an additional layer of beurocracy- potentially causing rows within government. Furthermore, would ordinary members of the public be pro-active enough to use the service? Would the small businessmen, burdened by regulation, have time?
By creating an ombudsman, you would be asking the public to make use of him. to be pro-active.
It is our job to be pro active. it is our duty to hold socialists to account, so people can work themselves towards a better tommorow.
Posted by: Matthew Barker | January 09, 2007 at 03:03 PM
If the proposed Statistical Ombudsman turned out to be as objective as the actual Parliamentary Ombudsman reported by Christopher Booker in Sunday's "Telegraph", all the dubious statistics of government spin would be given a seal of approval.
Far better (and cheaper) to believe the reverse of any official pronouncement, which is what all sensible persons do anyway.
Posted by: Father Brian | January 09, 2007 at 03:24 PM
One of the more sensible policy ideas I have seen on this forum though, like Wellness, I vaguely recall the Government announcing some sort of independent statistical body.
The downside, from the Conservative party point of view, is that well-presented statistics might undermine some of the campaigning slogans.
For example, I was always a bit puzzled by the '66 tax increases' which remained a constant Conservative refrain for a couple of years, suggesting the cessation of further money-grabbing activity by the Chancellor -- Then I heard Oliver Letwin speak at CCHQ and state "actually, there are now a lot more than 66 tax increases but it's a nice easy number to remember".
I suspect that much of the campainging on the climate change issue would founder in the face of sound statistical evidence.
To quote the late, great, David Ogilivy "some people use statistics like a drunk uses a lamp-post--for support rather than illumination".
An excellent idea though so best of luck with it.
Posted by: Les | January 09, 2007 at 03:31 PM
Great idea - go with it. This is the type of simple, but catchy idea we need, not like the expenisve, intrusive and impractical idea we had yesterday.
Posted by: Jack Bains | January 09, 2007 at 04:47 PM
Actually George Osborne was calling for the very important concept of Independence for the Government Statistical Service last year, and the ombudsman concept sits well with that. Remember though that while statistical evidence is objective about the thing its measuring, the inference one places on it is always subjective, so it's not a fault of statistics per se that different people come to different conclusions when faced with the same data. But I'm quibbling! Very sensible idea that would improve the governance of data - a good fit with freedom of information too by the way.
Posted by: Graeme Archer | January 10, 2007 at 01:08 PM
It's nice to see that my suggestion has been read and considered.
The government move to give the Office of National Statistics independance is too limited. They did report that the police and health figures for road casualities differed but did not go on to do the work necessary to decide how to get the most reliable figure. I suggest that peer pressure in the two organisations ensure that they both bias the figures, because the health service can claim costs on motoring related work whereas the police are incentivised to reduce accidents. Grey cases will thus tend to dealt with reflecting these pressures. It is not that individuals set out to cheat, just human nature to please the boss.
The main aim of the proposal is to stop management jumping to conclussions that are not supported by legitimate analysis.
John Allen
Posted by: John Allen | January 10, 2007 at 06:31 PM
Somewhat behind the curve:
Statistics and Registration Service Bill
Posted by: William Norton | January 15, 2007 at 05:55 PM
In theory, an ombudsman is an independent judge whose fair and impartial. In practice, they're political appointees often in the pockets of the players in the industry which they're supposed to regulate.
The European Parliament's recent enquiry in to how Equitable Life investors were treated by the Financial Ombudsman (as well as the FSA and HM Treasury) tell you all that you need to know about ombudsmen.
http://www.europarl.europa.eu/comparl/tempcom/equi/default_en.htm
Posted by: James Dey | May 31, 2007 at 12:56 PM