[News] Arrest after shooting in St Leonard's.

[looney]

No. They are not to be used as any sort of indication. The ONS made that very clear. No amount of sophism or pedantry regarding the definitions of invalid or unreliable will alter that. You want the quote verbatim? Here's the quote verbatim:

“While these figures are useful in giving an insight into the caseload of the police and how this is changing, they are not believed to provide a reliable measure of trends in violent crime,” the ONS report said. The ONS said that police-registered crime must be interpreted with caution, attributing much of the rise to changes in recording practices and increased confidence of victims in coming forward. Its preferred measure, the Crime Survey for England and Wales, gave a total of 10.6 million incidents, which was a fall of 10 per cent.

If you still believe, after that quote, that you can use the police figures as an indication of crime trends then you're saying that the ONS are wrong and that they do not understand statistics.

No because when they said......
"While these figures are useful in giving an insight into the caseload of the police and how this is changing, they are not believed to provide a reliable measure of trends in violent crime,"

They are saying the same as what I said. I said you paraphrased invalid when they were unreliable, or not beleived to be a reliable measure in their words.

And when they say,

"attributing much of the rise to changes in recording practices and increased confidence of victims in coming forward"

Attributing is conjecture, as is indicative which I said. OK attributing causes is different but its saying the same thing.

 

[looney]

Well he obviously gets them better than you do. He at least knows how to critique source and not to isolate data points.

Example?

See my reply to Buzzer.

 

[Buzzer]

No because when they said......
"While these figures are useful in giving an insight into the caseload of the police and how this is changing, they are not believed to provide a reliable measure of trends in violent crime,"

They are saying the same as what I said. I said you paraphrased invalid when they were unreliable, or not beleived to be a reliable measure in their word

And when they say,

"attributing much of the rise to changes in recording practices and increased confidence of victims in coming forward"

Attributing is conjecture, as is indicative which I said. OK attributing causes is different but its saying the same thing.

No, no and once again no.

Firstly, I did not substitute the word unreliable for invalid. I said the source data was unreliable for measuring trends and then said that any conclusions drawn from it about trends are invalid. It's inference, standard logic.

You on the other hand are taking confirmation bias to a stratospheric level in trying to argue that when the ONS say that the figures are an insight into the case-load of the police they are saying it's indicative of crime trends. I don't need to try to prove this isn't true because you've even quoted the relevant bit: "they are not believed to provide a reliable measure of trends in violent crime".

That bit about conjecture and attributing is just nonsense and all so much sophistry. It boils down to the ONS saying you can't use that 42% as a reliable measure of trends and you saying "yes, you can".

 

[looney]

No, no and once again no.

Firstly, I did not substitute the word unreliable for invalid. I said the source data was unreliable for measuring trends and then said that any conclusions drawn from it about trends are invalid. It's inference, standard logic.

Crossed wires, I agree thats what you said but the conclusion drawn from the trends is not invalid, ie falsified, the conclusion is unreliable.

That bit about conjecture and attributing is just nonsense and all so much sophistry. It boils down to the ONS saying you can't use that 42% as a reliable measure of trends and you saying "yes, you can".

Your just being silly now. I agree with what the ONS apart from the atribution of causes as this was their conjecture, I disagree with your using the term invalidate as it doesnt.

The BBC seem happy enough with the 42%, an organisation I have no confidence in and would likely play down something like this due to its ethnic dimension.

http://www.bbc.co.uk/news/uk-england-london-39578500

The assumptions you would need to validate the grey area in the stats along your lines are both that the 42%, or near to it was caused by an increase in shooting with airrifles, Even though hedonic/proxy measures for gun crime show a large rise, for example, the use of Londons air ambulance for gunshot wounds has overtaken call outs for road aciidents. Unless your claiming the invention of a full bore air rifle.


Secondly the argument that more people reporting certain crimes seems unlikely that people would become more motivated to report air rifle crimes than usual than hand gun crimes.

I think in this case rather than a statistics manual a dictionary would be of more help to you.

 

[Buzzer]

I disagree with your using the term invalidate as it doesnt.

I'm sorry but it really does. There's something called 'statistical conclusion validity' and the threats to this validity are:

I've highlighted the relevant bit. As I've said earlier, I'm not saying that the police data or even that increase of 42% is wrong, it may well be spot on but if it's based on data that the experts consider unreliable when assessing trends (%age increases/%age decreases) then any conclusions about trends might be accurate, might be wildly off - we just won't know for sure and therefore have to be disregarded.

 

[Surrey_Albion]

Terrible news,, very sad but applie named policeman Mark ROSSER

 

[looney]

I'm sorry but it really does. There's something called 'statistical conclusion validity' and the threats to this validity are:

I've highlighted the relevant bit. As I've said earlier, I'm not saying that the police data or even that increase of 42% is wrong, it may well be spot on but if it's based on data that the experts consider unreliable when assessing trends (%age increases/%age decreases) then any conclusions about trends might be accurate, might be wildly off - we just won't know for sure and therefore have to be disregarded.

No they dont, what you are doing is a type 1 error, kind of, you are rejecting a theory based on lack of evidence that may turn out to be true. You can not dismiss it as invalid through lack of evidence,(Thats a confirmation bias), you can dismiss it by testing the alternate hypothesis. The part about over or under estimating the size of the relationship between variables doesn't take place because there is no statistical estimate just a conjecture.

You are also using a logical fallacy with an appeal to authority argument. the ONS does make mistakes. In this case its not their numbers I am questioning its the assumptions made, attributed to the rise in crime.

 

[Springal]

Interesting to see he robbed a gun shop of the gun and ammo.... looks like they might be facing a bit of trouble now !

 

[Buzzer]

No they dont, what you are doing is a type 1 error, kind of, you are rejecting a theory based on lack of evidence that may turn out to be true. You can not dismiss it as invalid through lack of evidence,(Thats a confirmation bias), you can dismiss it by testing the alternate hypothesis. The part about over or under estimating the size of the relationship between variables doesn't take place because there is no statistical estimate just a conjecture.

You are also using a logical fallacy with an appeal to authority argument. the ONS does make mistakes. In this case its not their numbers I am questioning its the assumptions made, attributed to the rise in crime.

Sigh... . There really is no fallacy. I've already acknowledged that the figures may be right but because they are unreliable according to the ONS, I'd follow the rule on statistical analysis validity and ignore any conclusions. That is not confirmation bias. If you choose to believe otherwise then so be it. I'm all done here.

 

[Lady Whistledown]

Interesting to see he robbed a gun shop of the gun and ammo.... looks like they might be facing a bit of trouble now !

Well. The term "robbery" rather implies- in fact absolutely determines- a theft where force was used, rather than somebody just going "here you go buddy", so...

 

[Springal]

Well. The term "robbery" rather implies- in fact absolutely determines- a theft where force was used, rather than somebody just going "here you go buddy", so...

Oh yeah I get that but surely if you are a shop selling guns and ammo you'd hope it might take a bit of force to steal that kind of thing?

 

[looney]

Sigh... . There really is no fallacy. I've already acknowledged that the figures may be right but because they are unreliable according to the ONS, I'd follow the rule on statistical analysis validity and ignore any conclusions. That is not confirmation bias. If you choose to believe otherwise then so be it. I'm all done here.

I agree, there has to be more interesting things to argue about.