FDR or FWER with DIFFERENT tests

hello everyone,

My question is about correcting p-values for multiple testing, either through False Discovery Rate or FamilyWise Error Rate methods.

Can these methods be applied on a set of different tests (e.g. a mix of t-tests, Chi-square tests) ?



I would assume that you can mix your tests, and mix parametric and non-parametric ones, with FDR. I am looking for an example on the Web of such mixing. Anyone got any definitive answers? Thanks.