Yes I know what the narrative is, but I look at the facts.

There are, basically, three ways of measuring crime.

The Crime Survey of England and Wales
A&E figures for violent crime victims.
Police figures.

The first one, the CSEW checks lots of crimes which are never reported to the police. It is an expertly conducted, highly accurate picture of crimes actually experienced by people. It was set up after serious flaws in police recording of crime came to light. It finds crime continuing to fall.

A&E figures for violence requiring hospital treatment mirrors the findings of the CSEW. Clearly hospitals have no figures for other crimes, but the fact that it confirms the CSEWs findings for those areas it HAS knowledge of adds credence to the CSEWs figures.

The police figures are known to be of very poor quality. So bad that they are specifically excluded as a basis for policymaking. It is very noticeable that when cost-cutting is likely, the police response is to diligently record every crime. Under less threatening conditions they tend to discourage reporting of crime in order to reduce their 'unsolved crimes' list.

Yet we seem to have saturation coverage of the flaky police figures, with embarrassing figures from more accurate sources relegated to page 94 or not even mentioned.

If you want an example of fake news, then look no further than the publicity afforded to these blatantly fiddle figures.

Does it matter? Yes, it does. Policing is very expensive, and if it is unnecessary it means something else will suffer. the NHS, Schools Hospitals, Roads to the detriment of us all.

Figures for crime levels and trends for England and Wales are primarily on two sets of statistics. The Crime Survey for England and Wales (CSEW) and police recorded crime data. (CSEW) is a market survey of perceived crime and are not used as a stand alone measure of crime.

I still don't know why you have such faith in Kantar, you only have to look at their website which is over two years out of date in places to realise how competent they are http://www.crimesurvey.co.uk/HomeReadMore.html/. Just to prove it, the second page I went to was out of date as well http://www.crimesurvey.co.uk/SurveyResults.html

I'd be interested to know how they correlate the number of "incidents", surely the response to the survey would show the number of incidents multiplied by the number of witnesses? The victim based results are also hazy whether this is household or individual based.

I scan through and cross check a couple of figures to see how they map out, in a few minutes I spotted some suspicious figures ...

Theft offences has a lot more victims than incidents ( 4,399,000 victims, 3,311,000 incidents).

Apparently about 47% of adults have a bicycle, a figure I would have thought maybe pushing it at one time but seems less likely these days.

Statements like 80% of the population were not victims of crime is unrealistic when some of those could have been standing beside someone who had been glassed - yes, there is only one "victim" but there are many more affected. I don't now if the "5 Cap" limit of number of crimes still exists in the survey, something they promised to address for years.

The number of people 'glassed' is falling. (Hospital A&E figures)

This is at least partial corroboration of CSEW's figures. If you want to know the details of CSEWs methodology, then go and look at their website. the data. It is all there. I can't see why the theft statistics give you cause for concern. Someone steals your family car. One crime, but are both you and your wife not victims?

The fact that crime is falling everywhere in the west makes it likely that the same is true of us too. (One very plausible theory is that crime was high due to the use of TetraEthyl Lead in petrol. Countries who abandon its use, experience a dramatic fall in crime about 20 years later. We in the UK are seeing it it over the same timescale.)

What I find amazing is your total disbelief of CSEWs figures whilst they are completely transparent about their methods and your complete acceptance of the police's figures despite the fact that they are known to be so suspect they have been denied 'National Statistics' status. I am not really accusing the police of 'fiddling' but clearly a policeman who is worried about being made redundant is going to make every effort to justify his existence by recording EVERYTHING as a crime.

I would expect to find that the CSEW - being independent of such concerns - would produce results much freer of bias.

It is worth remembering that it is coming up to budget time, and all sorts of organisations are attempting to get more cash out of the government. I notice the armed forces for instance issuing well-publicised claims for more money to protect us from cyber attacks on our critical national infrastructure from Russia. Amusingly, they have evidently forgotten that only a week or two back, one of their generals was claiming that the self-same Russians were going to cut the underwater cables over which such attacks would be mounted!

Hopefully, our politicians are used to these tricks an will look at these claims through narrowed lids, although given the lot we have in now I am not counting on it.

I don't accept the police figures either, I've had more than enough dealings with them to realise how they work. But I think it would be easier to audit the police figures.

It is almost impossible to audit the CSEW figures end to end. Their source data is anonymised and also the source data is probably not very repeatable and cannot be cross-referenced to anything.

The police source data is not anonymised, is highly repeatable and can be cross-referenced, a simple sample check can be done.

I don't accept where there is money involved that there is independence especially when that money comes from the customer. The rule is to keep the customer happy and you can guarantee the Home Office has an agenda, their priorities are not just based on crime figures.

How would a smashed bus stop glass appear in the CSEW figures, no victim, no household and usually no perpetrator. Yes, if you read through all the CSEW text you may find its intentionally excluded but incidents like these are a major crime against society.

How about gangs of rowdy people (including kids) sitting on public steps using drugs, smashing bottles etc (yep, Egremont prom!).

How about kids openly nicking lifebelts (yep, Egremont Prom again).

The overall CSEW figures are not representative of the day to day crime that is highly visible and affects a lot of people. Funnily enough they are probably more representative of the figures that they state as less accurate because of the low numbers.

If you report incidents to the police they get an incident number (even if its only there to collect dust), the stats are in the system and can be extracted. If the police then convince you to withdraw the complaint, those stats are there as well. Its all traceable and hence auditable.

The point of the CSEW is that it covers crimes which never gets to the police such as groups of kids smashing bottles, people having their bins nicked, petty vandalism etc.Until the recent sudden boost in police figures, it showed a far larger rate of crime than the police claimed, albeit their figure was falling.

I cannot see why any polling company has any motivation to 'please the customer' by fiddling its results. The survey is commissioned by civil servants who have nothing to gain from a positive or a negative result. You are ignoring the findings of the A&E violent crime figures too. Are they also in on the conspiracy? And finally how come crime is falling in all western countries including ours - except by the police who are clearly worried about their funding being cut?

You seem to think the police figure 'can be audited' but they are not going to be. The data they produce is simply not good enough to take seriously.

I have in the past been discouraged by the police from requesting a crime number for a car theft on the basis that they would not be able to do anything anyway. That was at a time when they were keen to keep their list of 'unsolved crimes' as small as possible. I suspect now I'd get a crime number whether I wanted it or not, indeed, I might even be offered two or three! How you audit that I cannot imagine and clearly neither can anyone else.

I daily see several criminal offences that dont make up the crime stats...drivers driving using mobile. 3 just this morning in half an hour and thats the ones i spotted.

No incident number no insurance claim, you choosing not to take a copy of the incident number is irrelevant it still exists. A quango is a quango is a Quango. The questionnaire is 129 pages of stupidity "Don't include stolen milk bottles" mentioned several times says it all. Fact- one bike stolen or ten bikes stolen at the same time = one theft/ incident (CSEW) questionnaire.
Quango An organisation that is established by a government to consider a subject of public importance but is independent from the government.

The point of the CSEW is that it IS possible to pick up this sorts of event,Fish. Interviewees are asked if they have seen criminal activity

But I think the important fact is that the CSEW is seeing a steady downward trend using the same methods of assessing each year. If it only catches 50% of criminal activity, a fall of five or ten percent is still statistically significant. The survey involves around 40 or 50 thousand people being interviewed. This is an enormous number, when you realise that political pollsters claim 5% accuracy after interviewing only about a thousand people.

I return to my main point, however, which is that the media are primarily concerned with a 'good' story, and rising crime is a much 'better' story than falling crime. "Dog bites man!" is a story. "Dog doesn't bite man!" isn't. Similarly, politicians are very keen to be seen as protecting us from crime and seize every opportunity to promote that image. Both have picked out the story that best serves their purpose, and the result is a lot of people are needlessly terrified and much-needed resources are dissipated in unnecessary policing.