Sunday, March 31, 2013

Correlation does not imply causation. Or does it?


In a regular basis, we make conclusions about what we see or experience making the assumption that A happened before B, so A causes B. 

One, if not the first thing we discussed in class, was the fallacy of correlation proving causation.  The fact that there is a correlation between data might imply there is a relationship between two variables. A correlation might tell us the if a relationship is perfect, imperfect, negative or positive, yet, it does not tell us the causes of the relationship.  

Robert Henderson, the creator of Pastafarism or followers of the Flying spaghetti monster came up with a funny way of mocking this fallacy. According to this religious ideology, many factors such as global warming, earthquakes, tsunamis and many other natural disasters are caused by the shrinking number of pirates since the eighteen century.   He even goes fatherly by showing data about Somalia, pointing out that the reason why Somalia has the lowest carbon emissions in the world, it’s because Somalia happens to  have the highest number of pirates in the world. 

Correlation is not causation; correlation is just a hint that two variables are related.  There might be a causation, yet, we would need another type of test such as the Grainger test to prove the correlation

References
http://web.cn.edu/kwheeler/logic_causation.html
http://en.wikipedia.org/wiki/Flying_Spaghetti_Monster

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.