Statistical inference for treatments versus a control

Peng, Jianan (2002) Statistical inference for treatments versus a control. Doctoral (PhD) thesis, Memorial University of Newfoundland.

[English] PDF (Migrated (PDF/A Conversion) from original format: (application/pdf)) - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (16MB) [English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. (Original Version)

Abstract

The treatments versus a control problem occurs in many scientific fields, with a major portion in medical research. Its primary goal is to determine if the response to one or more treatments differ from the response to a control or existing standard and if so, further to identify which treatments are better than the control. In many experiments, one often has a prior knowledge that the treatments are at least as effective as the control. It is well known that utilization of ordering information increases the efficiency of statistical inference procedures. The aim of this thesis is to develop some new statistical inference procedures for the problem by utilizing the prior information. -- In particular, simultaneous confidence lower bounds for the differences between treatment means and the control mean are considered. Efficient computation algorithms are proposed to obtain the optimal lower bounds between the best treatment mean and the control mean. Multiple contrast tests which take account of the prior knowledge play an important role in this thesis. -- Power studies via simulation compare the new proposed procedures with Dunnett's procedure and the likelihood ratio test. The new proposed procedures are also illustrated by some real data sets.

Actions (login required)

View Item