[R] Test for stochastic dominance, non-inferiority test for distributions

[Matthias Gondan]

Dear R-Users,

Is anyone aware of a significance test which allows 
demonstrating that one distribution dominates another?

Let F(t) and G(t) be two distribution functions, the
alternative hypothesis would be something like:

F(t) >= G(t), for all t

null hypothesis: F(t) < G(t), for some t.


Best wishes,

Matthias


PS. This one would be ok, as well:

F(t) > G(t), for all t

null hypothesis: F(t) <= G(t), for some t.

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!