42 strange things that correlate:
http://tylervigen.com/
Obviously, it's interesting fuel for the "correlation is not causation" discussion, particularly because it's interesting to think about what the lurking or confounding variables might be.
What I also think is interesting about these graphs is some of the graphs that seem to follow each other closely, but don't really have that high of a correlation coefficient. For example, Number people who drowned by falling into a swimming-pool vs. Number of films Niclas Cage appeared in. Most of the data has an r above .9, which is good, but I think it would be interesting for kids to talk about why the curves on that graph seem to rise and fall together, but the correlation coefficient is not really that convincing of there being a statistical correlation.
Also cool: if you click on one of the variables, you can see how it correlates with a whole mess of other variables. This site could clearly could be a huge time suck for stats teacher trying to find interesting data to work from.
No comments:
Post a Comment