ePrints

Effect of prior specification on Bayesian design for two-sample comparison of a binary outcome

Lookup NU author(s): Professor John MatthewsORCiD

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

Closed forms of the optimal Bayesian design for the estimation of the log odds-ratio from a two-sample experiment are obtained. The effect of various specifications of prior distributions are considered and plausible circumstances where unequal allocation is optimal are presented.

Publication metadata

Author(s): Matthews JNS

Publication type: Article

Publication status: Published

Journal: American Statistician

Year: 1999

Volume: 53

Issue: 3

Pages: 254-256

Print publication date: 01/08/1999

ISSN (print): 0003-1305

ISSN (electronic): 1537-2731

Publisher: American Statistical Association

Share