Distribution‐free R sample tests for the hypothesis of parallelism of response profiles

Abstract

In this paper two nonparametric testa are given for testing the nullhypothesis of parallelity of response curves in r earnplea. The first procedure is done by a permutation teat, whose practical applicability is ensured by a FORTRAN-subroutine available from the author. As the computational work and time grows rapidly with sample sizes, a Monte-Carlo-solution is optionally given by the same procedure.Alternatively an asymptotically distribution-free test bawd on quadratic forme is constructed for large sample situations.

The Explication of Implicit Team Knowledge and Its Supporting Effect on Team Processes and Technical Innovations

“…Moreover, exact permutation tests offer opportunities for testing even with small samples, where parametric asymptotic tests are not likely to be very robust. However, both exact and parametric testing procedures are asymptotically equivalent under conditions of normality and homogeneity of variance (see Pyhel, 1980).…”