Layman's Statistics query
Discussion
(I've posted this here because the trigger for this query is the Met Police's Crime Figures)
Watched a bit of a Steve Norris interview last night where he stated that serious crime had gone up at least 15% since Ken Livingstone came to power.
What possible justification can there be for measuring the success or otherwise of a particular policy unless the timeframe is 'right'.
E.g. (These figures are made up)
1997 Labour came to power. Unemployment 2m.
2001 Labour re-elected. Unemployment 1.5m
Credible statement: "Unemplyment has fallen 25% under this Government".
Yes?
No?
I *know* why they manipulate the figures and use dodgy data to give the impression that things are heading in the right direction. However, people don't believe the true stories when they have to contend with all of the spin.
What is the solution?
Watched a bit of a Steve Norris interview last night where he stated that serious crime had gone up at least 15% since Ken Livingstone came to power.
What possible justification can there be for measuring the success or otherwise of a particular policy unless the timeframe is 'right'.
E.g. (These figures are made up)
1997 Labour came to power. Unemployment 2m.
2001 Labour re-elected. Unemployment 1.5m
Credible statement: "Unemplyment has fallen 25% under this Government".
Yes?
No?
I *know* why they manipulate the figures and use dodgy data to give the impression that things are heading in the right direction. However, people don't believe the true stories when they have to contend with all of the spin.
What is the solution?
The solution as far as I see it.
Make the ONS a truly independant body.
Pass legislation that states that if anyone wants to make a statistical comparison over public services that it has to be properly vetted and ratified by the ONS.
Until such a thing happens politicians will be able to spin a product to market and that helps no one but themselves.
The current Labour government are possibly the worst in history for this particular transgression.
Make the ONS a truly independant body.
Pass legislation that states that if anyone wants to make a statistical comparison over public services that it has to be properly vetted and ratified by the ONS.
Until such a thing happens politicians will be able to spin a product to market and that helps no one but themselves.
The current Labour government are possibly the worst in history for this particular transgression.
Plotloss said:
The solution as far as I see it.
Make the ONS a truly independant body.
Pass legislation that states that if anyone wants to make a statistical comparison over public services that it has to be properly vetted and ratified by the ONS.
Until such a thing happens politicians will be able to spin a product to market and that helps no one but themselves.
The current Labour government are possibly the worst in history for this particular transgression.
Plotloss
I've been thinking along these lines also. The problem is that this type of thing is not seen as sexy or mainstream and there will probably be not a great deal of interest in such a proposal.
IMHO I think it depends on how someone approaches a problem. Any problem. Personally, I want to understand what the problem is and then devise a proper solution. Quick fixes/workarounds are great as long as the root cause of the problem is properly addressed.
It seems that we spend far too much time fixing (or trying to fix) the symptoms of problems rather than the actual causes.
Plotloss said:
The current Labour government are possibly the worst in history for this particular transgression.
What utter bullshit. Taking the unemployment figures as an example, since rospa has already mentioned them, between 1979 and 1987, the Tories changed the way the unemployment figures are counted 32 times. Surprise, surprise, 31 of them had the effect of reducing the figure. Note, they did not stop in 1987, it's just that I stopped counting.

"Lies, damn lies and statistics" - forget who said it but absolutely true.
To be properly understood any statistical measure requires its context and makeup to be known. This is never going to happen with 'soundbite statistics'.
And the answer is NOT to make ONS the gatekeeper - they have not exactly covered thmeselves with glory in the last 12 months. More to the point, do you want to pay more taxes so that they can have ten times their current staffing - thought not!
To be properly understood any statistical measure requires its context and makeup to be known. This is never going to happen with 'soundbite statistics'.
And the answer is NOT to make ONS the gatekeeper - they have not exactly covered thmeselves with glory in the last 12 months. More to the point, do you want to pay more taxes so that they can have ten times their current staffing - thought not!
jeffreyarcher said:
Plotloss said:
The current Labour government are possibly the worst in history for this particular transgression.
What utter bullshit. Taking the unemployment figures as an example, since rospa has already mentioned them, between 1979 and 1987, the Tories changed the way the unemployment figures are counted 32 times. Surprise, surprise, 31 of them had the effect of reducing the figure. Note, they did not stop in 1987, it's just that I stopped counting.
For the life of me I don't understand why the new Labout Government of 1997 didn't reform this area so that the true unemplyment figures were published.
This would make the new government look honest and the outgoing government dishonest.
rospa said:
For the life of me I don't understand why the new Labout Government of 1997 didn't reform this area so that the true unemplyment figures were published.
This would make the new government look honest and the outgoing government dishonest.
They did. They implemented the ILO measure of counting. Whilst that would not result in as high a figure as the pre-1979 U.K. method would have given, it did put the figure up considerably from what it would otherwise have been.
Don't forget that Gordon Brown kas taked himself with cutting 1000's of civil service jobs. This is the same chap who has employed 1000's more civil servants.
Little wonder that the unemployment figures are going downward.
If anyone thinks that we couldn't cut a massive amount of expenditure by sacking imcompetent people, reducing totally unnecessary bureacracy and re-engineering the key processes that all of us have to endure, then I'll eat my hat.
While I am vehemently opposed to ID cards, I do believe we could cut a huge amount of unnecessary hassle out of the system if we were to have a single database that housed basic information on us.
That way, when I move house, I have to notify a single organisation who will then inform all relevant
organisations of my new details. No longer will I have to write to Inland Revenue, Local Authority, Credit Card companies, Banks, etc. etc..
There must be 1000's of people employed by the state just to process chnage of personal details. Reform it. Reduce paperwork. Automate it. Sack them.
Another one that really gets on my goat is the fact that if you are claiming JSA (Job Seekers Allowance), you still have to go through another round of almost identical paperwork to claim Council Tax Benefit. Why? (Actually, I know why. It is because one is administered by national government and one by local government)
What we need is a fairly simple system to means test people and then if they pass, allow them access to the full benefits. It is in no-one's interest to have different criteria for different benefits. It makes it all too complicated, time-consuming and expensive.
Little wonder that the unemployment figures are going downward.
If anyone thinks that we couldn't cut a massive amount of expenditure by sacking imcompetent people, reducing totally unnecessary bureacracy and re-engineering the key processes that all of us have to endure, then I'll eat my hat.
While I am vehemently opposed to ID cards, I do believe we could cut a huge amount of unnecessary hassle out of the system if we were to have a single database that housed basic information on us.
That way, when I move house, I have to notify a single organisation who will then inform all relevant
organisations of my new details. No longer will I have to write to Inland Revenue, Local Authority, Credit Card companies, Banks, etc. etc..
There must be 1000's of people employed by the state just to process chnage of personal details. Reform it. Reduce paperwork. Automate it. Sack them.
Another one that really gets on my goat is the fact that if you are claiming JSA (Job Seekers Allowance), you still have to go through another round of almost identical paperwork to claim Council Tax Benefit. Why? (Actually, I know why. It is because one is administered by national government and one by local government)
What we need is a fairly simple system to means test people and then if they pass, allow them access to the full benefits. It is in no-one's interest to have different criteria for different benefits. It makes it all too complicated, time-consuming and expensive.
Would you not agree that the problem with statistics and percentages is the fact that they can be massaged to convey whatever message is appropriate for the messenger?
Using pay rises as an example, if it is declared that someone has had a 50% payrise, everyone says wow!
Wow probaly would be appropriate if the initial salary is say £100,000 per year because it means that the person will receive a further £50,000 a year. If onthe other hand a person is only receiving £50 per year then it equates to only a further £25.00 per year.
So on the one hand 50% seems a lot (and it probably is to be fair) but it is only impressive when the actual figures are stacked up.
So when it comes to Government statistics, the percentage rule only looks good when we know the formula being applied and the paramiters that they are working in.
Does that make sense?
Using pay rises as an example, if it is declared that someone has had a 50% payrise, everyone says wow!
Wow probaly would be appropriate if the initial salary is say £100,000 per year because it means that the person will receive a further £50,000 a year. If onthe other hand a person is only receiving £50 per year then it equates to only a further £25.00 per year.
So on the one hand 50% seems a lot (and it probably is to be fair) but it is only impressive when the actual figures are stacked up.
So when it comes to Government statistics, the percentage rule only looks good when we know the formula being applied and the paramiters that they are working in.
Does that make sense?
Gassing Station | Speed, Plod & the Law | Top of Page | What's New | My Stuff


