Direct correlation

[Linkway]

Theresa May has dismissed claims that an increase in police numbers will help solve knife crime, insisting there is “no direct correlation” between the two.

(The Guardian, newspaper website, UK)

Theresa May apparently means that the reduction in UK police numbers (due to cuts in government funding over the last several years) did not (and does not) lead to an increase in knife crime.

In my view, by saying there is "no direct correlation", she is actually saying that reduction in police numbers won't produce reduced knife crime, and that increases in police numbers won't produce increases in knife crime. But that is definitely not what she meant.

Her quoted statement has attracted lots of media attention, but taken literally should be unremarkable.

Or are there multiple (and opposite) meanings to "direct correlation"?

 

[VicNicSor]

Just from the quote provided, "direct correlation" means the larger the number of police officers, the quicker knife crime is solved. I don't see any other way to interpret that (of course we could logically derive other correlations from this particular correlation, but that wouldn't be what she meant by "the two" which are police numbers and solving knife crime). She dismisses that correlation.

 

[lingobingo]

My only criticism of that statement would be that the construction does not clearly indicate that one of the “the two” is police numbers, not “an increase in police numbers”. In other words, it’s not sufficiently clear that the allegedly non-existent direct correlation is between (a) police numbers and (b) knife crime.

 

[Linkway]

Thanks, but the problem is that the Prime Minister meant an inverse (negative) correlation - an increase in police numbers would not produce a reduction in knife crime.

"Direct correlation" means, in statistical contexts, a positive correlation - the two variables rise together, or fall together.

 

[manfy]

Or are there multiple (and opposite) meanings to "direct correlation"?

I see no real problem with that statement.
A correlation is a relationship between two or more variables. That relationship might be directly proportional, inversly proportional, exponential, etc. etc.
In normal everyday language, I understand 'direct correlation' as an obvious or immediate connection -- as opposed to a connection that works via several variables, e.g. increase in interest rates as a means of stimulating economy. There clearly is a connection between interest rates and economy, it's just not a direct one; it goes via consumer confidence, consumer spending, etc. etc. until it impacts the economy as a whole.

Theresa May has dismissed claims that an increase in police numbers will help solve knife crime, insisting there is “no direct correlation” between the two.

If I had to make fun of her, I'd start with her choice of 'will help solve knife crime' instead of 'will help prevent knife crime'.
So, she's saying a bigger policeforce will not get more work done. Why not??
(Well, actually it's the journalists who chose to make it sound that way. Fair enough.)

 

[Uncle Jack]

Thanks, but the problem is that the Prime Minister meant an inverse (negative) correlation - an increase in police numbers would not produce a reduction in knife crime.

"Direct correlation" means, in statistical contexts, a positive correlation - the two variables rise together, or fall together.

I would not trust either politicians or journalists to use mathematical and statistical terms accurately, but here the two things are "increase in police numbers" and "solve knife crime". If there is a correlation, you would expect it to be a positive one.