This is an internal function that performs various versions of the FDR procedure, but without the modification described in section 4 of our J of Climate paper.

fdr.master(pvals, qlevel = 0.05, method = "original")



(required): a vector of pvals on which to conduct the multiple testing


the proportion of false positives desired


one of 'original', the original method of Benjamini & Hochberg (1995), or 'general', the method of Benjamini & Yekutieli (2001), which requires no assumptions about the p-values, but which is much more conservative. We recommend 'original' for climatological data, and suspect it works well generally for spatial data.


NULL if no significant tests, or a vector of the indices of the significant tests

See also

Other FDR: fdrBasic(), fdr(), propAlt(), storey()


Chris Paciorek