Cover Page

REPEATED MEASUREMENTS AND CROSS-OVER DESIGNS

The Late Damaraju Raghavarao

Laura H. Carnell Professor

Department of Statistics

Temple University

Philadelphia, Pennsylvania

Lakshmi Padgett

Janssen Research & Development, LLC

Spring House, Pennsylvania

logo

Dedicated to the memory of my Dad,
Damaraju Raghavarao,
for his love of family, friends, students, and statistics

Preface

Repeated measurement designs and cross-over designs are considered synonyms by several researchers. However, we consider repeated measurements terminology to be appropriate for a setting where units do not receive different treatments in the course of the study, whereas cross-over designs consider a setting where units receive different treatments in the experiment. In cross-over designs, residual effects are not used in the model when there is a washout period between the change of treatments. Such designs are also called row–column designs and two-way elimination of heterogeneity designs. When washout periods are not used between the change of treatments, the model considers residual effects; such cases may be called cross-over designs with residual effects. These two classes of designs will be called cross-over designs without residual effects (CODWOR) and cross-over designs with residual effects (CODWR).

Repeated measurements analysis is usually covered in multivariate analysis books, CODWOR in standard design and analysis of experiments textbooks, and CODWR in specialized books on cross-over designs.

The literature on repeated measurements and cross-over designs has grown rapidly in the last few decades in a wide spectrum of research disciplines. We strongly suggest that researchers and experimenters should be familiar with the basic concepts in repeated measurements and cross-over designs. With this objective in mind, this monograph provides an extensive but not exhaustive coverage of several topics of interest. The necessary mathematical results are provided along with SAS version 9.2 programming lines and the needed output to draw inferences. The specialty of this work is to bring together useful contributions in repeated measurements, CODWR, and CODWOR. This book will be very useful for researchers and experimenters working on these topics. It will also be suitable as a graduate level textbook dealing with special topics on experimental designs.

Acknowledgments

I would like to express my gratitude to Sharada, Venkatrayudu, Rhandi, and Chris for their love and support in encouraging me to complete this project that I had started with my Dad.

Lakshmi Padgett

January 2014