David R. Henderson  

Correlation versus Causation

PRINT
Lip Service... Public Choice in Foreign Polic...

correlation.png

This is from here.

HT to Timothy Taylor.


Comments and Sharing


CATEGORIES: Economic Methods




COMMENTS (2 to date)
ThaomasH writes:

More research is needed. :)

I suggest a randomized longitudinal cohort control study using a difference in differences or regression discontinuity methodology.

jxrjxr writes:

To make things even more disheartening, causation doesn't necessarily imply correlation either. If you observe a perfect driver trying to keep a constant speed in a hilly area, you would observe no speed-vs-gas-pedal correlation nor any speed-vs-road-angle correlation; you would find a gas-pedal-vs-road-angle correlation though.

I think there's a name for this type of scenario - (Milton) Friedman's Thermostat, but I'm not sure how popular that term is.

Comments for this entry have been closed
Return to top