# How to calculate Grubb's critical value (for an outlier test)

#### luks

##### New Member
Dear all,

I'm trying to calculate Grubb's critical value to find outliers. I found a table with 100 values for alpha=0.05 http://www.sediment.uni-goettingen.de/staff/dunkl/software/pep-grubbs.pdf
Further, I found this formular to calculate G-crit: ((n-1)*t-crit/sqrt(n*(df+(t-crit)^2))
http://www.real-statistics.com/stud...generalized-extreme-studentized-deviate-test/

When i compare the results, I get with this formular, and the table, they don't fit. Does someone have a better formular I can use in excel or can someone say if the above mentioned formular is correct and the table is wrong?

I also tried to use the formular mentioned in wikipedia: https://en.wikipedia.org/wiki/Grubbs'_test_for_outliers
But I had the same issues calculating the "upper critical value of the t-distribution", which did not work for me.