Successive comparisons between ordered normal means using isotonic regression estimators

ABSTRACT

In this study we discuss multiple comparison procedures for checking differences among a sequence of normal means with ordered restriction. Lee and Spurrier (1995) proposed a multiple comparison procedure which tests the difference between two adjacent means using the difference of sample means. In this study we propose a multiple comparison procedure modifying Lee and Spurrier's (1995) procedure using isotonic regression estimators instead of sample means. We determine the critical value for pairwise comparisons for a specified significance level. Furthermore, we formulate the power of the test. Finally, we give some numerical examples regarding critical values and power of the test intended to compare our procedure with Lee and Spurrier's (1995) procedure.

KEYWORDS:

• Likelihood ratio test estimator
• Multivariate t-distribution
• Power comparison

MATHEMATICS SUBJECT CLASSIFICATION:

• 62H15
• 62F30

Acknowledgements

The author is deeply grateful to the reviewer and the editors for their valuable comments and suggestions.