In relation to the selection of the best treatment, confidence intervals or predictive intervals are constructed for multiple comparison with the best, or for comparing a treatment or the selected treatment with the best of the rest. The tightness of these statements are not clear, sometimes mis-specified. In this paper, the tightness of two simultaneous confidence intervals given by Hsu (1981, 1984) and the predictive interval given by Mengersen and Bofinger (1988) for the known and unknown variance cases will be studied. Copyright © 1991 Taylor & Francis Group, LLC. All rights reserved.
CitationYu, P. L. H., & Lam, K. (1991). Tightness of some confidence and predictive intervals related to selection. Communications in Statistics - Theory and Methods, 20(4), 1401-1408. doi: 10.1080/03610929108830573
- Confidence interval
- Predictive interval
- Unknown variances
- Least favourable configuration