Thursday, January 21, 2010

On Correlation and Crime

Alderman Wisneski thinks there is a correlation between alcohol sold at restaurants and crime. Maybe so, but I suspect the correlation is rather weak. On the other hand, I think there is a much stronger correlation between the following, and crime:

1. Being a male.
2. Being young, say between the ages of 15 and 35.
3. Being black.

Perhaps Alderman Wisneski, if he wants to be consistent, will propose punitive measures on the young, males, and black people who want to move into Racine.

I guess it is somehow OK to profile restaurant patrons but not OK to profile those people more strongly correlated with crime. A better idea in my view would be to get tough on the people with a 1 to 1 correlation with crime, ie criminals.

12 comments:

BradK said...

But it's too much fun to base correlations on "feelings" versus "statistics"....

Anonymous said...

But it was a "direct correlation"!!!!

Can anyone explain what an "indirect correlation" is . . .

BradK said...

indirect correlation - a correlation in which large values of one variable are associated with small values of the other; the correlation coefficient is between 0 and -1

I think that very clearly spells it out... ;)

Anonymous said...

That would be an inverse correlation -

BradK said...

I thought inverse correlations were when variable x moves one direction, variable y moves the other...

Are you saying the Interweb LIED to me??!!

Stupid interweb...

Anonymous said...

Actually that's what I thought you meant by high of one being associated with low of another. If there really is an "indirect correlation" I've not heard of it or dealt with it. I will pray to Obama for an answer . . . he will most assuredly guide me, negatively -

BradK said...

I'm not 100% sure but I think indirect correlations - extrapolated out to an extreme - are the basis of the "Butterfly Effect" component of Chaos Theory.

I may have to watch Jurassic Park again to bone up on my science.

HA!

"bone up"

kay said...

Denis you named the three main factors in the highest poverty rates in Racine, Black, age group and gender ( I add gender because very often females of very low income have children and receive at least some sort of government aid....although, I do think that males tend to commit different sorts of crimes than females). So, don't you think you should have at least added poverty to the list?

Denis Navratil said...

Kay, my list was not meant to be an exhaustive one. Some people drink and commit crimes, thus there is a correlation. Some play frisbee and commit crimes, thus a correlation. Correlation is different from causation. My point was to expose the flawed thinking used by Alderman Wisneski who seems to have confused the two. In the process, he is demonizing/profiling restaurant owners and patrons.

Caledonication said...

I would say there is a far greater correlation between politicians and crime.

Caledonication said...

Oh yes.. Hello Denis my friend.

Denis Navratil said...

Hi Caledon. Keep those comments coming.