cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
Thierry_S
Super User

Benjamini-Hochberg (FDR) exact calculation?

Hi JMP Community,

For as long as I can remember, I have used an approximation of the Benjamini-Hochberg FDR calculation to account for multiple testing:

 > On a sorted list of Nominal P Values (smaller to larger) : BH p Val = Nom P Val * N Row()/Row ()

However, I'm aware of some limitation with this approach:

  1. For large Nominal p Values, the formula described above returns values > 1 (not by much but it tells me that something is not right)
  2. For Nominal p Values that are equal (or very close for one another), the formula returns slightly different values which I think should not be the case

Hence, is there a better way to calculate the FDR?

 

Thank you for your help.

Best,

TS

 

Thierry R. Sornasse
1 ACCEPTED SOLUTION

Accepted Solutions
Thierry_S
Super User

Re: Benjamini-Hochberg (FDR) exact calculation?

Hi JMP Community,

With a bit of digging, I was able to find an Add-In developed by John Sall from JMP that calculates correctly the FDR p Value False Discovery Rate PValue 

Best,

TS

Thierry R. Sornasse

View solution in original post

1 REPLY 1
Thierry_S
Super User

Re: Benjamini-Hochberg (FDR) exact calculation?

Hi JMP Community,

With a bit of digging, I was able to find an Add-In developed by John Sall from JMP that calculates correctly the FDR p Value False Discovery Rate PValue 

Best,

TS

Thierry R. Sornasse