Strong Type I Error Control and Point Estimation in Two-Stage Adaptive Designs


Material Information

Strong Type I Error Control and Point Estimation in Two-Stage Adaptive Designs
Physical Description:
1 online resource (69 p.)
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Committee Chair:
Wu, Samuel S
Committee Co-Chair:
Hobert, James P
Committee Members:
Ghosh, Malay
Lu, Xiaomin


Subjects / Keywords:
Statistics -- Dissertations, Academic -- UF
Statistics thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


A main objective of clinical trials is to test and prove the superiority of one or more treatments against a control. Often the best of several treatments is selected in a first trial. Then, in an independent second trial, superiority of this treatment to a control is evaluated using a size $\alpha$ test. This dissertation investigates the effectiveness of combining these two trials using a two-stage adaptive procedure, which is expected to improve statistical efficiency. Suppose we want to compare $k$ treatments with a control and the experiment is conducted in two stages. In the first stage, subjects are assigned to all $k$ treatments and the control. After the first stage, only those treatments with large sample means continue to the second stage along with the control, assuming larger means indicate more effective treatment. After the second stage, we decide whether any treatments are better than the control. In this dissertation, we study the mean-value combination method and give criteria to control the type I error. Then we use simulation to compare the power of this testing method to that of some existing methods. In addition, we consider the problem of estimating the mean difference between the selected treatment and the control, assuming only one treatment advances to the second stage. The naive sample mean estimator is shown to be inadmissible under the LINEX loss function. Under $L_2$ loss function, we show that the naive estimator is not minimax.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Anqi Sun.
Thesis (Ph.D.)--University of Florida, 2011.
Adviser: Wu, Samuel S.
Co-adviser: Hobert, James P.

Record Information

Source Institution:
Rights Management:
Applicable rights reserved.
lcc - LD1780 2011
System ID:

This item is only available as the following downloads:

Full Text