Statistics and fiddling

[#1]

So, many folks here are frequently referring to official statistics to back up their side of the arguement and disprove someone elses view.
So I ask those same people what they think of the revelation that the police have been fiddling their statistics?
/coat

[#2]

to save me searching - references please

[#3]

This maybe http://www.bbc.co.uk/news/uk-25002927

[#4]

Quote from: Oldjim

to save me searching - references please

Rather a skeptical reply from you there OJ - considering you always have links to everything from news to research!

[#5]

Google... the king of the fiddlers. 

[#6]

Strangely nobody who refers to official statistics on this forum seems to have any view on this!
I get a funny feeling though that we'll soon see some official statistics being posted to back up something in some thread somewhere.. 

[#7]

I Guess it's how you interpret the official statistics, not uncommon for people to have a different view on the same set of figures to support their own opinion.

[#8]

Slightly off topic - well not much but relevant
I was discussing this on a walk today with a senior nurse at a local hospital - not in Dorset - and she was very bitter about the system in place in the NHS where the Trusts are paid or not paid depending on how you meet the various tick boxes. In her case the hospital trust operates a very much open book system for problems etc. and as such they have been clobbered by a reduction in funding whereas those hospital trusts which have hidden things to improve their results are paid extra.
The point I am making in a roundabout way is that whenever you introduce a system where the organisation is rewarded for achieving certain artificial targets then the systems set up within that organisation will be geared to achieving those results even where such systems and targets aren't in the best interests of the customers or indeed to smooth running of the organisation.
My experience in private industry although probably not entirely relevant was that either the operation as a whole suffered or the figures were fudged (I won't say fiddled as it wasn't as bad a that)

[#9]

My illustration of statistics;
15% of accidents on the road are caused by drink drivers. Therefore 85% of accidents are caused by sober, responsible, people.
We have to stop sober people driving on the roads!!
( figures only to illustrate my view of statistics and statisticians).  Liars the lot of them! 

[#10]

That's exactly the point, Doris that I was going to make. Statistics can be used to argue for & against anything. They do not on their own mean a great deal unless backed up with comprehensive evidence (in my opinion anyway)

[#11]

The example doris uses proves a point worth noting.
If 15% of accidents are caused by drink driving then the only other conclusion that can be made is that 85% were caused by other means!
To suggest otherwise is guessing.
In doing the analysis, the stats guy will have examined the data for ALL accidents so will know the make-up of the 85%.
All stats must pass the 'sensibility test' - how can what I am presenting be challenged?

[#12]

Quote from: doris

15% of accidents on the road are caused by drink drivers. Therefore 85% of accidents are caused by sober, responsible, people. We have to stop sober people driving on the roads!!

More a case of post hoc fallacy than fiddling the figures. Fiddling is the job of the civil servant. Fallacy is the job of the politician.
Gabe