Current UK "leak" issue (UK Politics)

[fifty5]

There's all sorts of fuss about the "leak" ("statement" seems to me more accurate) of "partial and selective" "unchecked" statistics about knife crime in target areas of the UK.

To me the most shocking thing is that the production of these statistics didn't involve checking as part of the original process. For years I taught students about Validation and Verification - and that no data entry/gathering process was adequate without both. ('Validation' is automatic checking that input could be accurate. 'Verification' is checking - usually by two different people entering the same data, with automatic checks that they are identical.)

Note that both processes occur BEFORE any processing - to produce results - is done at all.

And of course, any computer code to calculate anything, really should be checked and tested properly before it is used 'in anger'.

Perhaps exceptionally, it seems to me that this time it is the Civil Servants who are at fault, rather than the politicians.

Producing any answer - so that it could possibly be leaked/released - that isn't properly checked and tested seems (on the little I've heard so far) tantamount to an announcement that "I'm not competent to do my job."

IMHO, the National Statistics Office ought to take their own share of the blame (almost all of it) instead of blaming the politicians.

 

[rgraham666]

Hmmm. About a month ago someone on here claimed that the violent crime rate in the UK is eight times higher than it is in the States. They had the statistics, supposedly, to back it up.

They blamed it on the fact that people in the UK aren't allowed to arm themselves.

 

[fifty5]

I've read the same argument about passports (ages ago, maybe it now applies to ID cards?...) that the underworld has means to obtain forged passports, but the law abiding citizen doesn't.

Your point may be valid, but I suspect it's more to do with different ways of collating statistics. I have no evidence to cite. Do you?

But my issue was about proper data processing: if that was OK, this issue shouldn't - if I understand correctly - have arisen at all.

WTF?

 

[rgraham666]

I tried, but I don't have any of the raw data. And I know nothing about statistics.

But that statistic just seemed wrong. I'm sure if things were that bad in the UK the people on here from there would have said something.

I can't prove it though.

 

[voluptuary_manque]

I tried, but I don't have any of the raw data. And I know nothing about statistics.

But that statistic just seemed wrong. I'm sure if things were that bad in the UK the people on here from there would have said something.

I can't prove it though.

I believe it all tends to involve where you live. Violent crime in the U.S. is very localized among certain populations. It's tragic that our minorities are doing their best to exterminate each other while our "European" population has about the same crime rate as Canada. Much the same thing holds true in Britain. Why, people? Someone tell me why this is?

 

[rgraham666]

Extreme poverty?

 

[voluptuary_manque]

Extreme poverty?

Nope. That doesn't do it. I know those neighborhoods and they don't qualify as 'extreme poverty'. Just ask R. Richard.

 

[fifty5]

I tried, but I don't have any of the raw data. And I know nothing about statistics.

But that statistic just seemed wrong. I'm sure if things were that bad in the UK the people on here from there would have said something.

I can't prove it though.

I think that what VM said is relevant - locally, I know of areas where poverty and deprivation make abiding by the letter and/or the spirit of the law an issue way below survival and - there's a quotation from the Declaration of Rights that I can't remember right now, but it comes down to self-respect.

Despite that, I do feel personally safer now than in the not too distant past. A few times, I've managed to leave keys in either the car or the back door. I'm no longer surprised that they were there when I've gone back the next day. A decade or so back, I had lights stolen/vandalised from the car and the garage door was forced to give entry to steal from the house.

However, what I wanted this thread to address was the issue of professionalism in computerised data processing. First I cannot imagine any modern statistics not being based on computerised calculation. Next, I find it totally incompetent for calculated statistics not to involve adequate checking BEFORE the statistics are generated.

That the civil service blames politicians for publishing such results before they have been checked seems to me like an admission of incompetence: why the hell were unchecked statistics calculated at all?

 

[rgraham666]

That one's easy, in my opinion.

Those sort of statistics frighten people. They don't know how these figures are derived and so believe them.

And frightened people are easy to manipulate.

 

[fifty5]

That one's easy, in my opinion.

Those sort of statistics frighten people. They don't know how these figures are derived and so believe them.

And frightened people are easy to manipulate.

Back to Yorkshire Ripper days - I took flack from folk when I suggested that despite the Ripper, women were much safer than they feared.

Nevertheless, the issue I want people to address is data processing: calculation of unchecked statistics is incompetent. Quoting them is a much lesser offence.

What have you to say on that issue?

 

[incubus666]

Here is a blog compairing crime rates in the U.S. and G.B.

http://wheelgun.blogspot.com/2007/01/crime-in-uk-versus-crime-in-us.html


It does include other links that can lead to other links.

At least part of the problem is the Home Office does not report crimes against anyone under 16yrs.

A few years ago Harvard University undertook a study about gun ownership and crime. The researchers fully admitted they believed high gun ownership increased violent crime. When the study was completed they said they were surprised the the exact opposite was true.

Another study compaired Canadian and U.S. border areas and crime rates. Canada has very strict gun control laws and the U.S. has far less. Again the researchers were quite surprised to find that violent crime was very much the same no matter which side of the boarder you lived in.

I used to have links to these studies but my HD died about a year ago and ate everything.

Google is free so take a look.

 

[rgraham666]

Back to Yorkshire Ripper days - I took flack from folk when I suggested that despite the Ripper, women were much safer than they feared.

Nevertheless, the issue I want people to address is data processing: calculation of unchecked statistics is incompetent. Quoting them is a much lesser offence.

What have you to say on that issue?

I'd say it's bloody unethical myself.

 

[fifty5]

Here is a blog compairing crime rates in the U.S. and G.B.

http://wheelgun.blogspot.com/2007/01/crime-in-uk-versus-crime-in-us.html


It does include other links that can lead to other links.

At least part of the problem is the Home Office does not report crimes against anyone under 16yrs.

A few years ago Harvard University undertook a study about gun ownership and crime. The researchers fully admitted they believed high gun ownership increased violent crime. When the study was completed they said they were surprised the the exact opposite was true.

Which seems to support what I said before about the basis of the statistics - it depends what you are measuring.

But PLEASE - what I'm trying to get at here isn't any result, but the process.

UK Civil Servants are blaming politicians for releasing statistics before they have been checked. What I'm saying is that the civil servants are incompetent for producing unchecked statistics.

Doesn't anyone have anything to say that's relevant to that?