Design and Analysis of Experiments

I used Dr. Montgomery's book as a reference material for his course under Arizona State University's six sigma black belt program. The book starts with a good review of experimental design concepts and inferential statistics. However, I do find his other book, Statistical Process Control, to be a much better inferential stat reference than this one because it does not 'rush thru' the review. Moving to chapter three is where the genius of the author becomes apparent. His approach was to discuss the theoretical aspects of a particular topic, relate it to back inferential statistics, discuss special cases or related topics (like a journal review paper), present problem/s and then show/discuss the solution using either design expert or JMP and diagrams. I cannot find any other better approach. Some of the previous books that I read are kind of 'cookbooks' which shows the reader how to do stuff using some method or software but did not really explain mathematically why it must be done in a particular way. This book did it elegantly. One reviewer mentioned that it is not a good book for hand calculation, which is true, but who does hand calculation these days? I cannot imagine myself pulling a graphing paper and drawing a response surface graph after working on several pages of calculations.The mathematics is not really hard ... only basic arithmetic and some basic linear algebra is all you need. Integrals/differentials and PDE's are mentioned in the second chapter to describe some functions but that's about it. However, it is important to understand the concepts early and sequentially as the topics build upon each other. For my case, one chapter takes about six hours total to read. The exercises are reasonable and emphasize the important concepts. It took me about 30 min to answer a reasonably challenging question. I would suggest buying the solutions manual to reduce your heartache or erase any doubts although be aware that not all exercises have a solutions in the manual.

I'm actually in Montgomery's class at ASU. The book is a great companion to the already great lectures, and the practice problems are very helpful. The math and theory is explained in excruciating detail, if you're into that sort of thing. You'd be hard-pressed to find a finer tome about design of experiments and basic statistics.With that being said, each edition adds and omits some information. So the current edition is up-to-date with current methods and research, but if you wanted to get an older edition, you could probably scrimp by. Some of the practice problems are different between the editions as well (in part to capitalize on the new content and remove the old), but if you don't have assigned bookwork in your class you could definitely get away with it. Alternately, you could also get a much cheaper older edition as well as the current edition solutions manual for half the cost of the book (at the time I bought mine). Only throwing this last bit in there because I know what it's like to be a poor, overworked engineering student and this might help you if you need options. ;)

This is a good textbook if you're taking a course in engineering statistics. It is very thorough and goes farther into the theory than most people might want. As an engineer I think the theory is valuable because it is easy to improperly apply statistical analysis. Overall I think the book is clearly written and provides many real life examples. To some of the other reviewers comment though, real world means an example about a process, but Montgomery tends to be light on telling you how to do the actual analysis. If you're looking for a how to do Minitab (or any other stat software) this isn't for you. If you want to learn about statistical analysis and DOE and how to do it right, then it's a great text.Summary:-Clearly written-Plenty of examples, though purchase the solutions guide if you've been assigned the homework-Lots of theory-Lots of real world examples-Don't expect to learn an analysis software packageNOTE: If you want the basics of DOE and statistical analysis applied in a practical manner, I would highly recommend the freely available NIST/SEMATECH e-Handbook of Statistical Methods, http://www.itl.nist.gov/div898/handbook/

Overall a very well written book that goes into great detail about most common aspects of experimental design. The only issue I had with this book was the adjustment to its style of writing and presentation of the information earlier on in the course. The first couple chapters where a little bit hard to fully appreciate the mathematics behind the statistics which in turn made understanding the important parts of the experimental design a little difficult. However, once you understand how to pull out particular information such as how the actual equations of certain designs differ from one another as well as the assumptions and constraint differences between fixed and random effects models the book is very consistent and easy to follow. This was for a graduate level statistics course.

If you're reading this review, there's a good chance that you are currently enrolled in Dr. Montgomery's "Design of Engineering Experiments" class. Prior to starting this class, I didn't have substantial knowledge on developing and analyzing experiments (above and beyond what was taught in undergraduate and masters classes) - but this class has proven one of the best that I have taken during my PhD. The book relies heavily on the JMP statistical processing software, so it is beneficial to also have this program installed in order to check book problem answers, and follow along on some of the book examples. The book lays out everything in a straightforward way, and I'm sure would be a valuable tool to all scientists and those in academia who are looking for reference material or a refresher on designing and analyzing experiments.