Thank you Terry. Right now I can use comp() from survMisc package to do the 2-parameter version of F-H weighting. I think both SAS and stata offer the 2-parameter version, so just?thought it would be nice if survdiff() can have that option given it's standard package in R.? Thanks! John On Friday, February 16, 2018, 7:08:46 AM PST, Therneau, Terry M., Ph.D. <therneau at mayo.edu> wrote: You are correct that the survreg routine only supports 'rho' of the Fleming-Harrington G-rho tests.? This is a function of age -- I wrote the original code back when I was working with Tom (Fleming), and he was only using 1 parameter. ?Later he and David expanded the test to two parameters.? This might be only the second request for the feature in the 30+ years since that date.? ? Terry Therneau [[alternative HTML version deleted]]