Concise Handbook of Experimental Methods for the Behavioral and Biological Sciences. (Book Reviews)
Anderson-Cook, Christine M., Journal of the American Statistical Association
Jay E. GOULD. Boca Raton, FL: CRC Press, 2002. ISBN 0-8493-1104-7. xvi+430 pp. 579.95 (P).
This is an interesting book on "the big picture" of experiments, written by a psychology professor from the University of West Florida (not Stephen J. Gould, the late, renowned zoology and geology professor). The author intends this as a book for practitioners who design experiments, to guide them through a number of the critical issues that are involved in getting the right data to answer the right question. Perhaps one of the book's most revealing aspects is how broadly some other disciplines view the experimental process, and how narrowly the impact of statisticians is viewed in this large process. George Box (Box 1999) summarized one of the book's unintentional messages for statisticians with his comment relating to drug discovery, but which is very broadly applicable: "While statisticians are accepted by scientists as necessary for the testing of a new drug, their value in helping to design the long series of experiments that lead to the discovery of a new drug is less likely to be recognized." This boo k is a useful reference for statisticians to gain insight into how their role is perceived and to become acquainted with the terminology and thought processes common in the behavioral and biological sciences.
I am sure that in a consulting or collaborative setting, many of us have had an experience where the client demonstrates a misunderstanding of what it is that statisticians do. Quite often we are viewed as the magicians, who at the end of the experiment wave their magic wand at the data and make some sensible numbers and conclusions appear. I view this book as a wake-up call to us statisticians to continue to broaden our mandate, and to package ourselves as problem solvers involved in all aspects of the scientific method--including the messy aspects of collecting data, selecting factors to consider, and making strategic decisions to deal with the compromises often necessary to meet time and budget constraints. (For an excellent book that tackles this broader mandate from a statistician's perspective, see Box 2000.)
The book is divided into three major sections: The first section is very general and discusses the philosophy of science and how experimentation fits into the scientific method. The second section discusses research ethics and experimental control, whereas the third, "Design and Analysis of Experiments," most closely resembles the material traditionally included in a statistics book on experimental design. The first section includes chapters titled "Philosophy of Science--the Scientific Approach to Knowledge," "Forms of Scientific Observation and Research" "Steps in the Scientific Method," "Variables in Research Designs" and "Initial and Final Phases of Research. …