Summary
In this note we consider a method proposed by Tukey [6], for detecting ordered alternatives amongk treatments in a randomized block design. A rank test obtained by this method is shown to be equivalent to a generalized sign test. The test is easily motivated and is quick and simple to compute, however its efficiency properties make it unattractive except for relatively smallk.
Similar content being viewed by others
References
Hollander, M. (1967). Rank tests for randomized blocks when the alternatives have ana priori ordering,Ann. Math. Statist.,38, 867–877.
Page, E. B. (1963). Ordered hypotheses for multiple treatments: a significance test for linear ranks,J. Amer. Statist. Ass.,58, 216–230.
Pirie, W. R. (1974). Comparing rank tests for ordered alternatives in randomized blocks,Ann. Statist.,2, 374–382.
Pirie, W. R. and Hollander, M. (1972). A distribution-free normal scores test for ordered alternatives in the randomized block design,J. Amer. Statist. Ass.,67, 855–857.
Puri, M. L. and Sen, P. K. (1968). On Chernoff-Savage tests for ordered alternatives in randomized blocks,Ann. Math. Statist.,39, 967–972.
Tukey, J. W. (1957). Sums of random partitions of ranks,Ann. Math. Statist.,28, 987–992.
Author information
Authors and Affiliations
About this article
Cite this article
Pirie, W.R., Hollander, M. Note on a Tukey test for ordered alternatives. Ann Inst Stat Math 27, 521–523 (1975). https://doi.org/10.1007/BF02504669
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02504669