This book places planned experimentation as part of an overall quality improvement effort. The design and analysis of analytic (as opposed to enumerative) studies is presented. Graphical methods of analysis are emphasized for all the classical experimental patterns. The book starts with the principles of design of analytic studies and presents designs for situations one would encounter in improving quality. A chapter on designing quality into products and processes is included. The book's final chapter contains two detailed case studies.
This book, which addresses experimental design in the social sciences, was first published in 1951. The emphasis is on observation rather than physical measurement. The first part of the book addresses the logical issues in planning studies, while the second and third sections address the practical issues of implementation. This book is different than others on experimental design in its emphasis on scales rather than continuous data. There is a good chapter on scaling for questionnaires and the development of questionnaires. The chapter on writing reports is useful.
Statistics for Experimenters: Design, Innovation, and Discovery, 2nd Edition
This is a classic book in experimental design. The authors are all leaders in developing the methods of experimentation in the United States. The
book does not address the concept of analytic studies, and was written before Taguchi's work became popular. The book has an excellent introduction on the role of statistics in the learning process. The book is divided into four parts:
Part I-Comparing Two Treatments - focuses on sampling and statistical issues. Randomization and simple blocking.
Part II-Comparing Two or More Treatments - analysis of variance, blocking, two factor factorial design
Part III-Measuring the Effect of Variables - modeling, factorial designs at two levels, use of cube and reference distribution, fractional factorial designs.
Part IV-Building Models and Using Them - least squares regression analysis, response surface methods, model building, nested designs, time series models.