Confidence lower bounds for comparing several treatments with a control in monotone dose-response studies
LE3 .A278 2020
Bachelor of Science
Mathematics & Statistics
Dunnett’s procedure (1955) is the most commonly used method to construct simultaneous confidence lower bounds to compare several treatment means with a control mean. It is common to assume that the treatment means are monotone in many dose-response studies. In this thesis, we propose a new but simple procedure based on Dunnett’s procedure to take into account the monotone assumption. We study the expected gains of the new procedure over Dunnett’s procedure. We illustrate our new procedure with numerical examples
The author retains copyright in this thesis. Any substantial copying or any other actions that exceed fair dealing or other exceptions in the Copyright Act require the permission of the author.