The breakfast experiments are some simple language experiments that the linguist Mark Liberman describes in his blog. For the corpus linguistics class of the master I am currently enrolled in I had to do one of this simple experiments.
In 2013 The Economist published a map of the consumption of tea and coffe around the world:
The data, aggregate from Euromonitor, confirmed my common sense knowledge: the western world mostly drinks coffe, while the east drinks tea. With due exception: Britain and Argentina, for different reasons, both consume more tea (or maté) than coffee.
The research question in my breakfast experiment was the following: is there any correlation between this data and language use? Afterall, if people talk about coffee they will probably drink it. One way to quickly answer this question is using Google NGrams.
As you can see the query results correlates very well with the original data. For Spanish there is no possibility to separate Argentina from Spain (which drinks a lot of coffe). For Chinese: I sense the results might be influenced by the fact the NGrams allows query for simpliefied Chinese only, but further investigations would be necessary to asses this. For a breakfast experiment this is enough, no ?