An approximate Bayesian up-down method for estimating a percentage point on a dose-response curve

Hui Li and Robert Malkin

Journal of Applied Statistics, 2000, vol. 27, issue 5, 579-587

Abstract: While the up-down method for estimating a percentage point on a dose-response curve has received considerable attention, a general Bayesian solution to the up-down design and estimation has never been presented, probably due to its computational complexity both in design and use. This paper presents a theoretical approach for up-down experimental designs with unknown location and slope parameters, and a practical approach for their use. The simplex method is used to find the optimal starting dose level and step sizes that minimize the expected root mean square error for a fixed number of observations and a reduced number of step sizes. The Bayesian estimate is then approximated by a polynomial formula. The coefficients of the formula are also chosen using simplex minimization. Two example solutions are given with uniform-uniform and normal-gamma joint prior distributions, showing that the simplifying assumptions make the method far easier to use with only a marginal increase in expected root mean square error. We show how to adapt these prior distributions to a wide range of frequently encountered applications.

Date: 2000
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.tandfonline.com/doi/abs/10.1080/02664760050076416 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:japsta:v:27:y:2000:i:5:p:579-587

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/CJAS20

DOI: 10.1080/02664760050076416

Access Statistics for this article

Journal of Applied Statistics is currently edited by Robert Aykroyd

More articles in Journal of Applied Statistics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().