Exact simultaneous confidence intervals for a finite set of contrasts of three, four or five generally correlated normal means
Exact simultaneous confidence intervals for a finite set of contrasts of three, four or five generally correlated normal means
The construction of a set of simultaneous confidence intervals for any finite number of contrasts of pp generally correlated normal means is considered. It is shown that the simultaneous confidence level can be expressed as a (p?2)(p?2)-dimensional integral for a general p?3p?3. This expression allows one to compute quickly and accurately, by using numerical quadrature, the required critical constants and multiplicity adjusted pp-values for at least p=3p=3, 4 and 5, involving only one-, two- and three-dimensional integrals, respectively. Real data examples from a drug stability study and a dose response study are used to illustrate the method
141-148
Liu, W.
b64150aa-d935-4209-804d-24c1b97e024a
Ah-Kine, P.
2553d7c8-99b4-492f-95ef-43c68e109737
Bretz, F.
51270819-e491-4a72-a410-679d86231e64
Hayter, A.J.
55bd07a5-db1d-4d3d-8c87-b307485420d9
January 2013
Liu, W.
b64150aa-d935-4209-804d-24c1b97e024a
Ah-Kine, P.
2553d7c8-99b4-492f-95ef-43c68e109737
Bretz, F.
51270819-e491-4a72-a410-679d86231e64
Hayter, A.J.
55bd07a5-db1d-4d3d-8c87-b307485420d9
Liu, W., Ah-Kine, P., Bretz, F. and Hayter, A.J.
(2013)
Exact simultaneous confidence intervals for a finite set of contrasts of three, four or five generally correlated normal means.
Computational Statistics & Data Analysis, 57 (1), .
(doi:10.1016/j.csda.2012.06.007).
Abstract
The construction of a set of simultaneous confidence intervals for any finite number of contrasts of pp generally correlated normal means is considered. It is shown that the simultaneous confidence level can be expressed as a (p?2)(p?2)-dimensional integral for a general p?3p?3. This expression allows one to compute quickly and accurately, by using numerical quadrature, the required critical constants and multiplicity adjusted pp-values for at least p=3p=3, 4 and 5, involving only one-, two- and three-dimensional integrals, respectively. Real data examples from a drug stability study and a dose response study are used to illustrate the method
This record has no associated files available for download.
More information
Published date: January 2013
Organisations:
Statistical Sciences Research Institute
Identifiers
Local EPrints ID: 361946
URI: http://eprints.soton.ac.uk/id/eprint/361946
ISSN: 0167-9473
PURE UUID: c2e38368-7562-475c-b0a0-6aa6eb162072
Catalogue record
Date deposited: 07 Feb 2014 10:26
Last modified: 15 Mar 2024 02:43
Export record
Altmetrics
Contributors
Author:
P. Ah-Kine
Author:
F. Bretz
Author:
A.J. Hayter
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics