More chocolate = more Nobel Prizes

Originally published at:

I’d like someone to point out to me what is epistemologically invalid with this line of reasoning, according to the scientific method. There is a hypothesis, accumulation of evidence, and a model, complete with a statistical test of goodness-of-fit.

if you think I am being facetious, I invite you to go read pretty much any other scientific paper out there today, which operates on more-or-less the same principles. If you think this epistemology is rubbish (and I do), you ought to have a huge, huge problem with the way science is operating right now (and I do). There are entire fields composed of papers exactly like this, where “the evidence at hands supports our suppositions” is the main method for producing new knowledge.


I don’t know, but you’d better send me all the chocolate you have to be sure. It may be our only hope!


Science doesn’t work on that principle. Correlation is not causation.


LOL depends what you mean by “work” and “principle”… there are so many different stages of scientific endeavor; from theory to experiment to measurement. and for many fields the “find anything that fits a pattern” is still what the field is trying to do.

so much of the working models science has built to explain complex phenomena are built from little or big correlations… and in the absence of good theory (e.g. biology), or perturbative analysis (e.g. economics), or both (e.g. sociology, psychology, etc), correlations are still the thing people turn to.

1 Like

There’s nothing wrong with the graph. Honestly I wouldn’t have a problem if a journal published this figure, as long as the associated paper didn’t make unjustified conclusions. The flaw is in assuming causation.

The next study might be seeing if the winning scientists ate more chocolate than their otherwise similarly prestigious peers, or whether scientists eat more chocolate than scientists, or a whole bunch of other things that together would quickly show the correlation was not causal. Science as an institution (human and flawed by nature) is organized so as to get the right answer eventually, not right away in a single study, even when doing so is possible.

This is also why popular reporting on science is so infuriating. Many/most scientists have a decent sense of how well they know things, but reporters are on the other side of the Dunning-Krueger divide.


I know this is done tongue in cheek to show that correlation is not causation but i would say that there is likely more truth in the inverse. Countries that produce cacao are less likely to have a population that has the means and education to produce Nobel laureates. A lot of the cacao bean production is not far from back breaking manual labor with not much profit for themselves. Countries that have the luxury of being able to more freely eat quality chocolate has the means to afford good education.


I wouldn’t say correlation is useless, especially when corroborated by further evidence. It can give you give you a reason to believe you may be on to something for further study. But you can’t draw a conclusion with correlation alone.


My favorite examples are Mars meteorites.

This rock came from the planet Mars, because it has the same chemical composition as rocks found on Mars! It’s science!

Sweden gets significantly more Nobel prizes than it’s chocolate consumption predicts. Who awards the Nobel prizes? I sense a scandal brewing…


Well, this is a surprise. Guess I’d better put down this chocolate bar and start writing that acceptance speech !

The only good way to test this hypothesis is to cut off chocolate supplies from those high consumption countries for 20 years and test the effect. Did the number of Nobel prizes decrease as chocolate consumption went down?

They could then send the excess chocolate to me, or perhaps by not consuming as much, increase the supply and lower the price.


Spurious correlations? Keep telling yourself that, but I eat tons of the stuff and it makes me a genius!


One of the researchers I did my masters with had a collection of articles like this. Horrible articles published in highly respected medical journals. There are some doosies. He would occasionally pull them out then he would say something about doctors needing to take epidemiology courses. This was one of the first ones he showed me.


What kind of choclate?

What IS causation, then?

You need to check the facts to avoid mixing up correlation and coincidence.
For example, the birth rate in the south of France and the size of the population of white storks (Ciconia ciconia) in the same area seem to be correlated at first glance.
But this is obviously rubbish as in France babies are not brought by the stork.
They grow in cabbage heads.


On first thought, I would have said the Nobel laureate ratio seems to correlate quite a bit with a decent educational system, but then again Austria is up there too and I have been through that one.

Ah well, that is a very good question, leading right into philosophy of science territory.

It is one of those concepts where you think you know quite well what it is, until you take a closer look…