How the Guinness Brewery Invented the Most Important Statistical Method in Science
Original Blog Post: How the Guinness Brewery Invented the Most Important Statistical Method in Science
Authors: Jack Murtagh (from Scientific American)
Published: 2024-05-25
My summary:
What does beer have to do with statistical significance?
In this article from Scientific American, Jack Murtagh discusses the origins of one of the most common tests of statistical significance, the t-test.
Around the start of the 20th century, Guinness wanted to improve the consistency of their products as they continued to expand worldwide. So, they hired a team to experiment and research how to improve their products.
One issue that kept coming up was small sample sizes. In other words, if they are assessing the quality of hops, instead of testing every flower in a crop, they wanted to test a small, random sample.
Therefore, the head experimental brewer at Guinness, William Sealy Gosset, invented the t-test.
If you need a review of the t-test and P-values, with some illustrations, take a look at the article.
Fun fact: Gosset published under the pseudonym “Student”, and therefore, the t-test is often referred to as the Student’s t-test, even today.
Here is the link again to the article for more details: How the Guinness Brewery Invented the Most Important Statistical Method in Science