Thursday, May 5, 2016

Correlation Does Not Imply Causation... But Wouldn't it Be Funny if it Did?


Does correlation imply causation? I think everyone and their grandmother at this point know that this is not the case. The assumption that correlation proves causation is incorrect logic -- just because two events occur together does not mean that they have a cause-and-effect relationship. There's even a name for this type of logical fallacy: "cum hoc ergo propter hoc," which is later for "with this, therefore because of this." However, allowing correlation to imply causation is just so so tempting. Scientifically, we correct for this by randomizing subjects and, ideally, being blinded until data analysis is complete.

Yet, if we were to conclude that correlation does indeed imply cause causation, how fun and interesting would the world be? We could blame the per capita consumption of margarine in the US for the falling divorce rate in Maine -- the two correlate to 99% according to the US Census and USDA.



Furthermore (with what I believe to be a 100% valid correlation) we could lower the number of people who drown by falling into a swimming pool in a very simple manner -- stop allowing Nicolas Cage to appear in films!



There would be some "innocent" (OR ARE THEY?!) victims though -- skiing facility revenue may decrease as people realize that funding these establishments directly correlated to a larger amount of people dying due to being tangled in their bed sheets.



And to think -- I could even validate my own consumption of mozzarella cheese! "There are simply so many Civil Engineering doctorates being awarded these days," I could tell myself. "I have no choice but to consume this cheese!"




No comments:

Post a Comment