For randomized block designs, nonparametric treatment comparisonsare usually made using the Friedman test for complete designs, and the Durbin testfor incomplete designs; see, for example, Conover (1998). This permits assessmentof only the mean rankings. Such comparisons are here extended to permit assessmentsof bivariate effects such as the linear by linear effect and the quadratic bylinear, or umbrella effect.
展开▼