Friday at 3:00 pm brings engineers and scientists at Crowd Science together to sit down with a good craft beer and chat. When we are finished, we then get back to work writing code.
We push no code on Friday afternoons, and therefore have no fear of blowing p-value calculations due to our mini-parties. In fact, we have a historical precedent for our consumption, because most statisticians know the link between statistics and beer, specifically, Guinness.
Guinness hired chemist William Gosset to apply statistics to the production of beer. In 1908 Gosset created and published what is now called the Student’s t-test to find differences in batches of stout.
The t-test allowed Guinness to maintain high standards, while signaling to chemists when the batches had gone awry.
Crowd Scientists wait weeks or even months to collect enough data to pass similar tests in what we collect for our customers. Once we collect sufficient data to pass our own tests, publishers, agencies and advertisers then base important ad buying decisions on this this information. It is after these key milestones we then celebrate with a brewski in Gosset’s honor.