Degree Type

Dissertation

Date of Award

1994

Degree Name

Doctor of Philosophy

Department

Statistics

First Advisor

Herbert T. David

Abstract

This research aims to enhance the efficiency of two-stage plans, by allowing second-stage sample size and critical region to depend on first-stage evidence. This aim is implemented by certain optimizations. These optimizations are done under two types of formulation; one formulation is in a sense related to the group-sequential point of view, while the other has simultaneously a Bayes and Neyman-Pearson interpretation. By prior choice of type of optimization, the experimenter can, to some extent, pre-determine the behavior of second-stage sample size as a function of first-stage outcome for the optimal plan. Both formulations show a modest reduction in overall sampling effort, as compared with matched optimal standard two-stage plans;When the two types of formulation are merged, there results a methodology capable of easily constructing whole families of approximately optimal binomial plans, based on a single Wiener optimization.

DOI

https://doi.org/10.31274/rtd-180813-9743

Publisher

Digital Repository @ Iowa State University, http://lib.dr.iastate.edu/

Copyright Owner

Seoung-gon Ko

Language

en

Proquest ID

AAI9503575

File Format

application/pdf

File Size

112 pages

Share

COinS