Hacker News Reader

Comment by gnosis | Hacker News Reader

There was a very interesting article[1] recently about the epidemiological link between leaded gasoline and crime.

[1] - http://www.motherjones.com/environment/2013/01/lead-crime-li...

I honestly find it dismaying that the Economist could publish a story specifically on the recent drop in crime and never have heard of this study.

I just read this story and its longer follow-up in the print edition. They do mention the lead study.

Then I find it dismaying that they know of it but intentionally refrained from mentioning it in the online edition. More dismaying, honestly, as they are specifically championing a panopticon solution to crime here and saying that increased police expenditure and profiling has resulted in lower crime rates.

The economist generally has 1 page "summary" articles at the front of the newspaper, and more in depth articles in the dedicated sections. Both are available online. Trying to put every point into the summary would defeat the entire purpose.

Except that the link between the drop in crime and the elimination of lead has very strong evidence, and the various social theories bandied about by the Economist have almost no support. They didn't just not put "every point" into the summary -- they completely ignored the biggest, best supported reason.

Here's the longer article: http://www.economist.com/news/briefing/21582041-rich-world-s...

The economist clearly doesn't agree with you, here's what they said about lead:

Jessica Wolpaw Reyes, an American researcher, has argued that the cognitive effects of exposure to lead were a primary determinant of violent crime, and unleaded petrol is to thank for the improvement. But the causal link is far from proven.

The Economist needs to read up a bit. There is strong support for lead as the primary determinant, and not just correlation.

http://www.motherjones.com/environment/2013/01/lead-crime-li...

Next time you feel like criticizing something you should probably know more than epsilon about it...

Well, you've certainly made a great point there, swordswinger12!

Correlation != Causation?

Correlation isn't causation, but there are methods you can use to get at the question of whether there is a causal link somewhere, and these methods involve looking at some correlations.

Reply on news.ycombinator.com