A Talk by Mitchel Fruin
Senior Data Scientist,
DWP Digital
About this talk
In the past year, the Universal Credit Data Science team at DWP piloted the use of a Bayesian estimation approach to analyse their A/B tests. This talk will explore the limitations of traditional Frequentist null hypothesis significance testing that motivated them to try a new approach, the benefits and difficulties of the Bayesian estimation approach they adopted, and an example use case.