Design and analysis of cross-over trials. (English) Zbl 0729.62068

Monographs on Statistics and Applied Probability, 34. London etc.: Chapman & Hall. xii, 340 p. $ 49.95 (1989).
This is the first book to give a full-length treatment of cross-over designs. These designs are widely used in medical research, psychology, dairy science, agriculture and other diverse areas. The literature is widely scattered in both theoretical and applied journals. The authors have done an excellent job in putting together this monograph with up-to- date literature. This text will be a useful reference book for practitioners of cross-over trials.
Chapter 1 gives a brief introduction of the topic. Chapters 2 and 3 discuss the commonly used two-treatment design in two periods. Methods of analyzing quantitative responses using both parametric and nonparametric techniques, the role of baseline measurements and Bayesian analysis are discussed in Chapter 2. Methods of analyzing binary and categorical data using simple contingency table procedures and log-linear models are given in Chapter 3. Other two-treatment designs with more sequences and/or more periods are presented in Chapter 4. A discussion of which two-treatment design to use, given in this chapter, will be helpful to the user. More mathematically motivated designs for three or more treatments and their ramifications form the content of Chapter 5. The analysis of the designs using repeated measurements and with covariance structures for the observations is given in Chapters 6 and 7.
The book is practically free from typographic errors and mistakes except that credit should be given to BR, for Blaisdell and Raghavarao, instead of PR in Table 5.16 and subsequently BR should be mentioned instead of PR on page 223.


62K99 Design of statistical experiments
62-02 Research exposition (monographs, survey articles) pertaining to statistics