Simultaneous Confidence Intervals for Monotone Contrasts of Normal Means
Authors:
Ruth Marcus
Affiliation:
Department of Statistics, University of Haifa
Abstract:
The one-way ANOVA model is considered. The variances are assumed to be either equal but unknown or unequal and unknown. Two-stage procedures for generating simultaneous confidence intervals (SCI) for the class of monotone contrasts of the means are presented. The intervals are of fixed length and independent of the unknown variances.