Statistical Design and Analysis of Experiments

Readers will find this book an invaluable reference on the design of experiments. It contains hard-to-find information on topics such as change-over designs with residual effects and early treatment of analysis of covariance. Other topics include linear models and quadratic forms, experiments with one or more factors, Latin square designs, and fractions of 2n factorial designs. There is also extensive coverage of the analysis of incomplete block designs and of the existence and construction of balanced and partially balanced designs. A new preface (to the classics edition) describes the changes made in experimental design since the book was first published in 1971. It discusses the use of personal computers to analyze data and details the emergence of industrial statistics.

'Peter John's book is truly a classic but not an outdated one. It was one of the first to blend in the standard matrix notation that almost all linear models books now use. It has topics and discussion that are still valuable today.' Richard F. Gunst, Southern Methodist University