# Grubb test

To use the Grubb test, we initially calculate:
- X: The average of the entirety of the measurements
- S: The standard deviation of the entirety of the measurements
- G: Distance between the value suspected of being an abnormal point and the average 
![equationInline](G=|"Value"-bar(X)|/S)
 
![grubb](baseUrl/statsDescriptives/grubb/Grubb-2.png)

Next, we compare the G value that was found with a G limit: :
![equation](G_{"limit"}=(N-1)/sqrt(N)*sqrt(t_{alpha/N,N-2)^2/(N-2+t_{alpha/N,N-2)^2)

If G > G<sub>limite</sub> then the value is considered abnormal and vice versa

User Guide

Grubb test