Fundamental Concepts in the Design of Experiments, 5/e offers comprehensive coverage of the key elements of experimental design used by applied researchers to solve problems in the field. Wide-ranging and accessible, it shows students how to use applied statistics for planning, running, and analyzing experiments. Featuring over 350 problems taken from the authors' actual industrial consulting experiences, the text gives students valuable practice with real data and problem solving. The problems emphasize the basic philosophy of design and are simple enough for students with limited mathematical backgrounds to understand. The authors provide extensive coverage of the analysis of residuals, the concept of resolution in fractional replications, Plackett-Burman designs, and Taguchi techniques. SAS (Statistical Analysis System) computer programs are incorporated to facilitate analysis. Thoroughly revised and updated, this new edition includes sixty new problems, focuses more on computer use (adding computer outputs from statistical packages like Minitab, SPSS, and JMP), and emphasizes graphical procedures including residual plots and normal quantile plots. Ideal for various advanced undergraduate and graduate experimental methods courses taught in statistics, engineering, and mathematics departments, this book will also appeal to professionals and researchers doing experimental work.
This was a very unclear introduction for an experimental-design novice like myself, though once I bashed my head against a wall long enough to get an understanding of the basics, the later chapters were actually pretty well written and easy to follow. It just took me FOREVER to see any kind of pattern in this topic... At first experimental design just seems like a haphazard jumble of heuristics without rhyme or reason, before you can see the overall structure of the subject, and this book wasn't as helpful as it should have been in getting me there. The book also had lots of good examples, but many were not followed through all the way. And there was no mathematical justification for some of the key concepts.