Whether this is a joke or not (I was seriously contemplating writing a letter to the editor describing its stupidity), it serves as an awesome example of poor judgement, statistical misuse, and confounds. I have given it to my multivariate statistics students to practice writing critical annotations. In the meantime, I thought about another predictor for Nobel prizes. Since the number of Nobel prizes won by residents of a country has nothing to do with national chocolate consumption, it should have a strong relationship to the resources made available by a country to furthering science in every direction. Countries with such priorities typically have residents with more disposable income than countries that do not. And thus we find more money is spent on indulgences like chocolate.
I replicated the experiment, and confirm that the correlation between Nobel prize per capita and chocolate consumption per capita is strongly correlated, with a correlation coefficient of 0.625. I fit the best quadratic to the data as well, from which we see a diminishing utility of chocolate dosage to Nobel prize winning.
But what about something more ridiculous?
So I looked at how the per capita Nobel prizes of a country is correlated with the percentage of a country on facebook. And indeed it is! With a correlation coefficient of 0.564. But more remarkably, the Nobel prize winning capacity of a country exponentially increases with the percentage of the country on facebook. Note that 0.946 is the starting per capita Nobel prize winning of a country without facebook, just because at least one person occupies that country.
So, today I am going to encourage my students to peruse facebook during the rest of my lectures because we have proven statistically the power of facebook for their scientific development. Our prizes await.