Students Tasting Chocolate – A Different Take on Fisher's Exact Test

Can you tell the difference between American and European chocolate? Students in Cornell Stats Professor Melissa Smith’s Categorical Data Analysis class can, having tested their tastebuds in a delicious lesson on Ronald Fisher’s exact test earlier this semester. 

The story goes that Fisher first experimented with his statistical significance test on a woman who claimed she could sip her tea and detect whether the milk or the tea was first poured into her cup. This became known as the “Lady Tasting Tea” experiment. The Data Analysis class substituted the tea for American and European brands of chocolate.

How did the students fare? Very well: All three groups did better than had they randomly guessed, and one team correctly identified each of the dozen chocolate pieces.