Tables for testing ordered alternatives in an analysis of variance

Abstract

In an analysis of variance of a one-way classification, the null hypothesis customarily states that the samples have been drawn from populations having the same mean. One useful alternative hypothesis orders the means in a monotone increasing sequence. Tables for testing this hypothesis are given for k means each having n observations for k = 3(1)10, n = 2(1)20,24,30,40,60, α = 0⋅10, 0⋅05, 0⋅01, 0⋅001.