On Jun 13, 1:34 am, Ray Koopman <koop...@sfu.ca> wrote: > On Jun 12, 8:12 am, Fijoy <tofi...@gmail.com> wrote: > > > > > > > Hi all, > > > I have two samples, s1 and s2, of axial data (set of unit vectors, > > where sign of each vector is irrelevant). The two samples are paired, > > i.e. each vector in s1 corresponds exactly to one vector in s2. > > > I want to test the null hypothesis that the means of the two samples > > are equal. That is, I am looking for something that is similar to the > > paired t-test on axial distributions. Does anyone know of such a test? > > > I do have the book "Directional Statistics" by Kanti Mardia and Peter > > Jupp, and the book has a test for equality of means on axial > > distributions, but this test in the book is an unpaired (indepedent) > > version. > > > Thank you for your help. > > > Sincerely, > > Fijoy Vadakkumpadan > > This is not an area that I know much about, but since no one else has > answered, let me guess that the solution will depend on the data only > through the squares of the inner products of corresponding s1 and s2 > vectors. Can you map that into an appropriate single-sample test?- Hide quoted text - > > - Show quoted text -
Thank you for the response. Off the top of my head, I suppose I could test the squares of inner products and see how close they are to zero. I am not sure how I can map this onto a statistical test. Anyways, I am trying to see first if I can directly deal with the unit vectors themselves instead of reducing them to a scalar parameter.