Premium
Robustness of choice of number of doses for maximum likelihood estimation of the ED 50 in bioassay
Author(s) -
Huang Yangxin
Publication year - 2002
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.1204
Subject(s) - robustness (evolution) , mathematics , statistics , probit model , mean squared error , probit , logistic regression , logit , biochemistry , chemistry , gene
A number of studies have looked at the choice of number of doses for estimating the median effective dose (ED 50 ) when the form of a dose–response curve is correctly assumed (see, for example, Müller and Schmitt, 1990). Here we address the question of the robustness of number of doses chosen using Müller and Schmitt's method when a logistic dose–response curve is incorrectly assumed. The underlying true dose–response curves considered include the probit, cubic‐logistic and Aranda‐Ordaz asymmetric models. The simulation results show that, for the above true dose–response curves and the uniform design density with doses spaced symmetrically or slightly asymmetrically around the assumed ED 50 , choosing as many doses as possible either minimizes the mean squared error (MSE) or produces a MSE chosen to the minimum. Copyright © 2002 John Wiley & Sons, Ltd.