Date of Award

2001

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Mathematics and Statistics

Abstract

A major obstacle in the analysis of experimental data, in many situations, is the lack of "true" or "complete" replication. In some disciplines, researchers are very aware of the importance of replication and the methods for correctly replicating an experiment. In other subject areas, however, researchers are less aware of what it means to properly replicate an experiment. Due to this lack of awareness, many non-replicated experiments are carried out every year. For many of these non-replicated experiments, there is no satisfactory statistical analysis.

The subject of this report is the analysis of two non-replicated experiments in environmental engineering. First, we analyze the data sets using two traditional methods that assume true replication. We then present an analysis that uses information from both experiments to analyze the individual experiments. We believe that this last analysis is superior to the traditional alternatives.

The organization of the rest of this report is as follows. In section two we summarize some background material on the analysis of factorial, split plot, and split-split plot experiments. In section three we introduce the data that were analyzed for this report. In section four we describe the three analyses we carried out. Finally, in section five, we compare and discuss the results of the three analyses.

Share

COinS