In a data analysis, does it make sense to report both Kendall's tau and Goodman's Gamma? I guess not, since tau looks like a standardized form of Gamma. Or does it add some explanation?
Asked
Active
Viewed 550 times
1
-
3If the data have no ties, gamma = tau. If the data have ties, you should think what to do with them - leave out? ignore? take into account - how? Hence you arrive at gamma or tau-a, tau-b, tau-c, etc. Please search this site for `Gamma`, `Kendall`. – ttnphns Jan 14 '14 at 11:58
-
1Make sure you happen across [ttnphns' answer to a related question](http://stats.stackexchange.com/a/18136/32036) in your search! – Nick Stauner Jan 14 '14 at 13:39