Heavy Lifting - thoughts and web finds by an economist
     I also contribute to Division of Labour Load HL's Front Page
Monday, August 23, 2004

Causation and Correlation (again)

The confusion between causation and correlation is rampant in our society.

In yet another example of why econometricians should review all statistics published in peer-reviewed journals in every discipline (okay, that was a bit of self aggrandizement), this article claims that
New research from the National Cancer Institute in Bethesda, Maryland shows a startling association between birth month and the risk of developing brain cancer in adulthood, reports Reuters.
The study was published in the journal Neurology (Abstract):
Results: A relationship between month of birth and risk of adult glioma and meningioma was found, best described by a 12-month periodic function with peaks in February and January and troughs in August and July. The association between month of birth and risk of glioma differed significantly by handedness, with left-handed and ambidextrous subjects born during late fall through early spring being at particularly high risk of adult glioma as compared with those born at other times of the year.

Whew. I was born in August, so I have a smaller risk of brain cancer? I haven't read the paper (UTA doesn't have a subscription), but if the statistics are what they seem to be, this is purely correlation and not causation. While the results might ultimately lead to future understanding of brain cancer, it is also possible that the results motivate expensive (both in terms of money and opportunity cost) research chasing down a dead-end.

Such results are more than difficult to publish in the economics literature. It is often claimed, mainly by econometricians and applied economists, that economics is fifteen to twenty years ahead of other fields in terms of empirical techniques - especially when it comes to determining causation. While econometrics doesn't get everything right all the time, it is not difficult to do better than confusing causation with correlation - and duping the media into the mistake as well.

Comments: Post a Comment


Le Chai - galerie du vin



Posts that contain Craig Depken per day for the last 90 days.


Heavy Lifting's Main Page
Email Me
Atom Feed

Heavy Lifting

Great Links

Money I Found Today

Heavy Lifting - Firehose style (56k warning)

Recent Posts

- Causation and Correlation (again)



Site Meter Blogroll Me!


Modified maystar design
powered by blogger