IB Mathematics
Organizer:
- Organized by
-
-
Correlation does not imply causation!
-
Correlation does not imply causation!
A very important concept in Statistics is the idea of correlation: we observe the behaviour or trend of two different variables (one independent, one dependent) and aim to decipher whether there is a mutual connection or common pattern between both. However, it is equally important to understand that the fact that two variables are correlated does not NECESSARILY mean that one (the independent) causally influences the other (the dependent). Even though it is true that a relevant correlation analysis focuses on two variables that have the POTENTIAL of being causally related (why would we bother in doing a correlation analysis otherwise?), a correlation coefficient does not suffice to establish such causal relationship; it can, at most, show the similarity in behaviour or trend between both variables and can serve as an indication that there MIGHT be a causal relationship. A correlation coefficient can however not conclusively establish whether a causal relationship exists.
I have attached a link to a website that shows very strong (and somewhat funny) correlations between variables that are clearly unrelated to each other. It turns out, for instance, that per capita cheese consumption in the US is strongly correlated (r= 0.9471) with the number of people who have died in the US by becoming tangled in their bed sheets (Data: 2000-2009). Even though per capita cheese consumption does not cause these uncommon sorts of deaths, there still exists a very strong correlation.
Hope you find this interesting and instructive! I will be posting more content like this in the future.
-
0 Replies
Sorry, there were no replies found.
Log in to reply.
