Effect of a Guarantee Time on Sample Size Determination for Testing the Ratio of Means from Two Lifetime Distributions

Chengjie Xiong; J. Philip Miller; Yan Yan; Ming Ji

doi:10.1081/BIP-120028513

Effect of a Guarantee Time on Sample Size Determination for Testing the Ratio of Means from Two Lifetime Distributions

Chengjie Xiong, J. Philip Miller, Yan Yan, Ming Ji

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

This paper studies the problem of sample size determination for testing the ratio of means from two lifetime distributions when thresholds, or guarantee times, are present in the survival distributions. We consider several different approaches of sample size determination based on different distributional assumptions and discuss the role that guarantee times play in these sample size computations. We show that the guarantee times, if they exist, can significantly affect the sample size computations. The Aging Intervention Testing Program from the National Institute on Aging is used to demonstrate our results.

Original language	English
Pages (from-to)	177-188
Number of pages	12
Journal	Journal of Biopharmaceutical Statistics
Volume	14
Issue number	1
DOIs	https://doi.org/10.1081/BIP-120028513
State	Published - 2004

Keywords

Guarantee time
Location-scale family
Lognormal distribution
Maximum likelihood estimate (MLE)
Proportional hazards
Type II censoring
Weibull distribution

Access to Document

10.1081/BIP-120028513

Cite this

@article{8984e258419245779a3a1b2318d072ee,

title = "Effect of a Guarantee Time on Sample Size Determination for Testing the Ratio of Means from Two Lifetime Distributions",

abstract = "This paper studies the problem of sample size determination for testing the ratio of means from two lifetime distributions when thresholds, or guarantee times, are present in the survival distributions. We consider several different approaches of sample size determination based on different distributional assumptions and discuss the role that guarantee times play in these sample size computations. We show that the guarantee times, if they exist, can significantly affect the sample size computations. The Aging Intervention Testing Program from the National Institute on Aging is used to demonstrate our results.",

keywords = "Guarantee time, Location-scale family, Lognormal distribution, Maximum likelihood estimate (MLE), Proportional hazards, Type II censoring, Weibull distribution",

author = "Chengjie Xiong and Miller, {J. Philip} and Yan Yan and Ming Ji",

note = "Funding Information: Dr. Xiong and Prof. Miller{\textquoteright}s work was partly supported by National Institute on Aging grants A6 03991 and A6 05681.",

year = "2004",

doi = "10.1081/BIP-120028513",

language = "English",

volume = "14",

pages = "177--188",

journal = "Journal of Biopharmaceutical Statistics",

issn = "1054-3406",

number = "1",

}

TY - JOUR

T1 - Effect of a Guarantee Time on Sample Size Determination for Testing the Ratio of Means from Two Lifetime Distributions

AU - Xiong, Chengjie

AU - Miller, J. Philip

AU - Yan, Yan

AU - Ji, Ming

N1 - Funding Information: Dr. Xiong and Prof. Miller’s work was partly supported by National Institute on Aging grants A6 03991 and A6 05681.

PY - 2004

Y1 - 2004

N2 - This paper studies the problem of sample size determination for testing the ratio of means from two lifetime distributions when thresholds, or guarantee times, are present in the survival distributions. We consider several different approaches of sample size determination based on different distributional assumptions and discuss the role that guarantee times play in these sample size computations. We show that the guarantee times, if they exist, can significantly affect the sample size computations. The Aging Intervention Testing Program from the National Institute on Aging is used to demonstrate our results.

AB - This paper studies the problem of sample size determination for testing the ratio of means from two lifetime distributions when thresholds, or guarantee times, are present in the survival distributions. We consider several different approaches of sample size determination based on different distributional assumptions and discuss the role that guarantee times play in these sample size computations. We show that the guarantee times, if they exist, can significantly affect the sample size computations. The Aging Intervention Testing Program from the National Institute on Aging is used to demonstrate our results.

KW - Guarantee time

KW - Location-scale family

KW - Lognormal distribution

KW - Maximum likelihood estimate (MLE)

KW - Proportional hazards

KW - Type II censoring

KW - Weibull distribution

UR - http://www.scopus.com/inward/record.url?scp=1542349173&partnerID=8YFLogxK

U2 - 10.1081/BIP-120028513

DO - 10.1081/BIP-120028513

M3 - Article

C2 - 15027507

AN - SCOPUS:1542349173

SN - 1054-3406

VL - 14

SP - 177

EP - 188

JO - Journal of Biopharmaceutical Statistics

JF - Journal of Biopharmaceutical Statistics

IS - 1

ER -

Effect of a Guarantee Time on Sample Size Determination for Testing the Ratio of Means from Two Lifetime Distributions

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this