What is the difference between FDR and FWE?

However, for a procedure to be good, we want it to make few errors. This is where FDR and FWE come in, they formalize the amount of errors. Suppose we denote R the total number of hypotheses that we called false (or rejected), and we call V the total number of errors made. read more

FDR, the false discovery rate, is theexpected value ratio of errors we have made: [math]E(V/max(R,1))[/math]. Note the funny-looking denominator in the ratio: we take the maximum of R and 1. This is simply to ensure that we do not divide by zero. read more

Difference in FWE and FDR thresholds between xjView and SPM8 Showing 1-8 of 8 messages read more