Toggle Main Menu Toggle Search

Open Access padlockePrints

Assurance for sample size determination in reliability demonstration testing

Lookup NU author(s): Dr Kevin Wilson, Dr Malcolm Farrow



This is the authors' accepted manuscript of an article that has been published in its final definitive form by Taylor & Francis, 2021.

For re-use rights please refer to the publisher's terms and conditions.


Manufacturers are required to demonstrate that products meet reliability targets. A wayto achieve this is with reliability demonstration tests (RDTs), where a number of products areput on test and the test is passed or failed according to a decision rule based on the observedoutcomes. There are various methods for determining the sample size for RDTs, typicallybased on the power of a hypothesis test following the RDT or risk criteria. Bayesian riskcriteria approaches combine the choice of sample size with the analysis of the test data whilerelying on the specification of acceptable and rejectable reliability levels. In this paper weoffer an alternative approach to sample size determination based on the idea of assurance. Thisapproach chooses the sample size to provide a specified probability that the RDT will result ina successful outcome. It separates the design and analysis of the RDT, allowing different priorsfor the producer and consumer. We develop the assurance approach for sample size calculationsin RDTs for binomial and Weibull likelihoods and propose appropriate prior distributions forthe design and analysis of the test. In each case, we illustrate the approach with an examplebased on real data. Supplementary materials for this article are available online.

Publication metadata

Author(s): Wilson KJ, Farrow M

Publication type: Article

Publication status: Published

Journal: Technometrics

Year: 2021

Volume: 63

Issue: 4

Pages: 523-535

Print publication date: 01/10/2021

Online publication date: 23/12/2020

Acceptance date: 05/12/2020

Date deposited: 07/12/2020

ISSN (print): 0040-1706

ISSN (electronic): 1537-2723

Publisher: Taylor & Francis


DOI: 10.1080/00401706.2020.1867646


Altmetrics provided by Altmetric