experimental-economics

George Box is one of the statistical giants of the 20th Century. He started his career in chemical engineering in England where he learned the importance of experimental design and statistical methods. He came to the US in 1953 and spent time at North Carolina State College and later came back to be part of the statistics group at Princeton. After that he founded the Department fo Statistics at the University of Wisconsin. This history and other important career decisions icluding the founding of Technometrics are detailed in the brief section "My Professional Life" that he wrote for this volume.

Box's contributions to statistics are diverse and large. He developed many practical statistical designs including the central composite design. He is responsible for evolutionary operation and wrote a book on it with Norman Draper. He has also made major contributions to response surface methodology.

With Gwilym Jenkins he systematized the application of the ARIMA models and led the development of software for easy application of these model building techniques. He championed the concept of parsimonious models and insisted that model building should be an iterative and continually evolving technique. He contributed to the area of control through his stochastic time series models and found ways to incorporate it in manufacturing process control.

With David Cox he developed the Box-Cox family of transformations. These simple power transformation can be used to make the data have an approximate normal shape. he gave a prescription for how to estimate or pick the power to use based on the data.

These enormous contributions can be found in the volumes of collected works that Tiao and others have edited. His contributions can also be seen from his books on evolutionary operation, time series analysis, automated process control, empirical model building and response surfaces, and practical experimental designs ("Statistics for Experimenters").

However in the decades of the 80s and 90s from age 60 to 80, instead of retiring, George Box took on the challenge of developing a center for quality and productivity at the University of Wisconsin. This volme, edited by Tiao, Bisgaard, Hill, Pena and Stigler provides a collection of articles by Box. These are mostly articles written in the 1990s covering the subjects of A) continuous process improvement, B) designing experiments to gain quality information, C) sequential investigation and discovery (including response surface methods), D) quality control and E) learning how to identify and reduce variation or be less sensitive to it by constructing robust processes (i.e. processes not sensitive to minor changes in process parameters). The articles are mostly directed toward quality issues and are mostly articles that were published in the 1990s or 2000 with a few from the 80s. Some are important technical contributions but many are also very philosophical.

George Box is one of the great thinkers of the 20th century and his philosophy on statistics and scientific inference is as important as his many technical contributions. There are 46 articles in total 4 on topic A, 12 on B, 10 on C, 11 on D and 9 on E. Each topic area has a brief introduction identifying a unifying theme in the papers in that section.

Box has a terrific sense of humor that often comes out in his lectures and sometimes in his writings. One gets a good appreciation of it by reading the three songs on statistics that are included in Part F of the book. This is only a sample of several that he has written that are parodies of familiar tunes. Of these three my favorite is "There's no theorem like Bayes theorem" to the tune of "There's no business like show business."

There is a nice bibliography in the back of the book that is followed by a biography on Box and a list of his books and articles published between 1982 and 1999. This includes 3 books and 91 articles! Believe it or not he published even more in his earlier years.

I recently used this book in an introductory course in experimental methods in political science, taught at Columbia University. (I used the book in a seminar for juniors but it would also be appropriate for first year graduate students.) The book was recommended to me by a leading experimentalist in political science, and its easy to see why. It is filled with practical suggestions, on everything from recruiting subjects, to writing instructions, to analyzing data. It's many references to the literature (as well as sample syllabi) are very helpful. It is also short and readable, and reasonably priced. One caveat: if you are more interested in the psychology tradition to laboratory experiments, there are more appropriate books. Also, it does not deal with field experiments. But if you are interested in the game theoretic/economic tradition of laboratory experiments, it is hard to beat as a practical introduction. In short, a terrific book. Charles Cameron Associate Professor of Political Science Columbia University

Who wants to know how experimental economics has evolved, what experimental economists cover and do? Anyone who has ever thought why people make seemingly irrational decisions can read this book and find out how the research has been evolving in that regard! Find interesting differences from economists' view, compared with psychologists' view. The book is a must for any experimental economists, and all researchers interested in judgment and decision making. This is a great book for any experimental economics course... and the website maintained by Alvin Roth is a treasure for further research.

This book is a great introduction to ballanced orthogonal experimental designs. The so-called Scientific Method (tweak one parameter while holding the other variables fixed) is not used that much where there are complicated interactions between variables. This Designed Experimental (DOX) approach is used more frequently in the Semiconductor industry. This approach envolves a series of experiments where all the variables are potentially changed form one trial to another, but in such a way that one can resolved the effects of one variable from another AND understand interactions and non-linear effects. It is simply the best way to optimize a complicated system.

OK, if you have SAS or RS/1 or ...packages you may ask yourself "why bother learning the theory." I would say for 2 reasons - if you need to manually intervene or supplement an existing experiment AND seeing some more designs that are very effecient (John's 3/4 designs and Plackett-Burman designs) for initial screening that is not ususally part of the packages.

Compared to Box, Box, Hunter this book is easier to read, has a huge summary fold out sheet and is geared to a practicing engineer or scientist not a professional mathematician. Best introduction and most likely the only book you will ever need on the subject.

This is the best quality text book I have ever read! I learned so much and was able to apply it to my job and my boss has really noticed my increased contribution! Thanks Thomas B. Barker!

Finally, teachers and administrators have a useful tool to recharge their classrooms/schools. We have become so test oriented that many times we forget that the process of learning comes from the doing. The statement comes to mind that we utter with new approaches," When do I have the time to get it all in?". That is where the problem lies! We are so pressured with these new demands that we loose sight of what is important, the process. This book will give us a clearer picture of how we can recharge our teaching and can provide the key so we can use the ideas to change the way we teach or approach the students' learning who are in our care.

I was getting ready to prepare a training program for 44 persons when I read this book. Immediately! after I used the guidelines to prepare my session. Preparation was fast and easy -and fun! I really couldn't believe that it was that easy to apply and it has been one of my best sessions to date. I refer to it when I need a refresher or a new perspective. I definately reccomend this book for trainers who need to keep their material fresh and interesting (esp. in-house trainers).

I am a subject matter expert who was occassionally requested to make training presentations. I began to realize that training was my calling. I was then hired to plan training programs. By chance, I ran across this book. After reading the book, I was so inspired I wanted to read it again to ensure that I didn't miss anything important. Here is what happened next. I needed to plan training for professionals in a country I had never visited -- truly a risky venture! Relying on the principles in this book, I designed a program that really involved the participants (the students). For example, in the afternoon of the first day of the workshop, students were meeting with each other in breakout sessions, identifying "promising practices" and "best practices" and making presentations to the whole group. The workshop lasted three days. My supervisors really liked the design of the program. The instructors/facilitators loved participating in the program. And, most importantly, the students/participants loved it so much that I am told to expect twice as many students/participants at the next session. As one instructor/facilitator put it, if the goal of the program was to both create joy and share knowledge, the goal was achieved. Without the information in this book (about how the brain works, about the importance of joy to successful training, about the importance of giving the students/participants the opportunity to create knowledge, etc., etc.), my program would not have been the success that it was. To the author -- Dave Meier (whom I have never met) -- I have these words: "Your book inspired me. It inspired me to design a great training program ... a training program even better than I could have imagined. Because of the importance of the training program, I will look back at the program as one of the great achievements in my life. And I owe it to you. Thank you."

I bought this book hoping to learn DOE using Taguchi Methods and apply what I learned to real life problems in my work. It was very helpful. Not only did it help me learn the methods I needed to perform my testing at work but I did it in only 8 days of reading. The book was very easy to read and the examples helped me understand everything I needed to know to apply this method to real situations. A great buy for anyone looking to learn Taguchi Methods.

The challenge that all research directors and senior scientists face is that virtually all scientists are originally trained to believe that the only proper way to conduct an experiment is to vary only one factor at a time while holding all other variables fixed. This belief, which can be shown mathematically to be totally incorrect, has been a drag on the scientific community for decades, causing experimental science to be much less productive than it should be. In fact, by properly designing experiments in which several factors are varied in a carefully thought out, predetermined pattern, much more information can be obtained with far fewer experiments.

What Dr. Roy has achieved in his book on experimental design is to clearly explain why this is so and to provide the tools which allows his readers to overcome their ingrained beliefs and adopt a more effective method of designing their own experimental programs.

Of particular interest is his discussion of experimental design using orthogonal arrays (Chapter 4) for it is in this part of the book that the full power of modern approaches to experimentation become most evident. Here, he illustrates how a set of as few as eight experiments can be used to determine how three presumable independent factors, such as time, temperature and concentration, can be studied to obtain detailed information not only on how each factor individually effects the quality of the final product, but also on the extent to which the factors interact.

Historically, learning this type of technique for designing experiments has been a daunting task since somewhat tedious mathematically manipulations are required for both the design of the experiment and the analyses of the resulting data. However, included with Dr. Roy's book is complete software which eliminates the need for the experimenter to either manually solve the handful of equations needed to extract the results in useful form, or to develop custom spreadsheets in an attempt to automate the process. The software is easy to use and includes all of the tables and data which are used in the book to illustrate the principles of experimental design. The software is also capable of assisting the experimenter in designing sets of up to eight concurrent experiments, rapidly analyzing the data and generating graphical and tabular presentations which greatly aid in the interpretation of the results.

This is an extremely useful book which can have a major beneficial effect on the productivity of any laboratory engaged in experimental process research such as crystal growth, chemical synthesis or manufacturing. By allowing the reader to overcome his innate aversion to varying more than one experimental factor at a time, the book makes it possible for the reader to become a much more productive scientist or engineer and become a role model for his coworkers to emulate.

First, I completely agree with all the good things mentioned in the other reviews, but wish to add the following comments:

1. The discussion group (usenet) hosted by Dr Roy is fantastic. If by chance you don't "get it" from reading the book, a simple question posted to the usenet frequently gets you an answer, often within a few hours (its a vocal crowd!!). Many times, the author himself responds to usenet questions. And, on at least one instance, after I repeatdly asked my question, he began exchanging emails with me, which then led to some pleasant phone disscussions.

2. User friendly software and support. See above.

3. Taguichi is frequently attacked by full blown Design of Experiments statistical types who miss the whole point. I strongly encourage all professionals interested in gauging quality control to try out his books and the free software downloads....

An excellent book with powerful software and support......

Reads like a winner, looks like a winner, performs likes a winner....... need I say more...

Auburn Alabama

I have used this book when it was in its first edition and only by Neter and Wasserman. An absolute must for those who use linear models (regression and ANOVA) and want the most exhaustive book out there. This book is one of the rare texts that is also extremely well written. You can not go wrong reading it and studying it on your own. If you can take only one statistics book with you, this is the one. I use the 4th edition so much, I wore my first one out and am now on the second.

This book seems very difficult and long to read at first, but as soon as I got into the habit of using this book, this is really a great book to have. The terminologies are very easy to understand and yet not redundant, and it has some great reference areas in the sections that you could look up for information when you need to work on an exercise or a project. The exercises in the back of each section are ample enough so that you can have a firm grasp of the material after working on those exercise. This book is useful and is a must have for any math/stat student!

Enjoy...

Wonderful explanation of regression methods with many worked examples. The book is very detailed and covers practically everything you need to know on the topic. It also comes with a disk containing the example datasets. I used it for years in graduate school and still use it today. If you work through the book, you can teach yourself all about regression - everything you need is there (although no free lunch - this isn't exactly easy material)

This book is for someone that has identified problems, and wants to use sound experiments to solve those problems. The book provides some good guidelines, and some good forms for preparing experiments (some of which I use). They emphasize graphical/visual approaches, and avoid statistics, which I missed somewhat. When I travel to different plants to solve problems, I use this book as a reference.

Robert Brewer has written a book that allows you to understand statistical process control techniques without being a mathematician. He uses graphs were ever possible to point out how to tell if a process is working as it should. He also explains ideas like process capability, process improvment and analysis of variance of properties in a straighforward, readable and understandable fashion.

this book makes statistics understandable. i like the graphics!

This is a very timely book in my opinion. Prensky gives a reasonable overview of trends in games and several likely future developments. Copious lists of heuristics and tips are offered to the reader. Almost all the case studies have at least one useful insight - however many case studies drag on too long and are in need of substantial editing. Moreover, at least 50% of the case studies relate to corporate programmes which are not accessible to the general public - or the research community.

Prensky's writing style is intensely personal and his judgements are often based on subjective assesments. Worst of all, he turns over large sections of the book to game designers that uncritically trumpet their products. There is still something of value here, once you read between the lines.

In summation I found the reviews of trends and products useful. His explanation for what keeps a game's audience and market together (content) resonates with my own experiences. The book fell down by being mistitled. It is not about game based learning but about games and the opportunities for game based learning. Do no expect to build a 'learning' game from what is here - just not academic enough. The book has little if any information on existing game engines and how they might accommodate learning initiatives which is a major shortcoming. Secondly, many of the major points in the book flow from anecdotes and opinions. It is hard to assess their worth, but experience can be a better teacher on occasions than a library. Thirdly, the book plugs Prensky's own work fairly relentlessly. Good for him but it creates an imbalance in the presentation.

Overall, I found much in the book that was interesting and useful to know. If the opinion pieces were supported by more complete referencing it would be an excellent text.

When you set out to tell all about e-learning games, it helps to have a broad background. Harvard MBA Marc Prensky has been a professional musician, high school teacher, business consultant, corporate trainer, and creator of dozens of software e-learning games. He is now CEO of a game company. He brings to the subject an intense love that enables him to find value even in flawed efforts in this fledgeling field, yet he retains the critical capacity and balanced judgment that are the hallmarks of credibility. The book provides a provocative analysis of how widespread playing of video and computer games has created a new under-35 Games Generation that sees the world through radically different eyes than their parents, who can be at best "digital immigrants." Simply transferring tell-test and "sage on the stage" teaching methods to digital media, as many e-learning companies have done, misses the point, he explains. We need to exploit the opportunities to develop new, more engaging approaches--and that primarily means games. People will learn the most technical or boring subjects if presented as part of compelling, fun games. Prensky offers scores of case studies and war stories from practitioners in the corporate trenches. He ranges from simple quiz games to intensive virtual reality simulations, showing that often simple is better, that not only the type of learner but also the kind of subject/skill to be mastered should determine the approach, and that games possess inherent advantages over simulations. He laughs at the outrageous scenarios players deliberately generate in customer relationship games. He notes with respect the capacity of deep simulations and games to draw upon the insights of communities of experts worldwide to generate cutting-edge research results that can then be parlayed into real-world management systems. And he states six objections to e-learning games, then proceeds to blow them to bits like so many villains in a shooter game. In the book are many practical suggestions for getting an e-learning game project approved and funded as well as for how to bring it into reality. Of special value are its list of state-of-the-art games in various categories and its articulation of the advantages of each type. Two suggestions for the second edition: 1) Games and e-learning are at the cutting edge of a controversial transformation of cultures around the world, so it would be nice to have a chapter on e-learning games outside the United States; and 2) someone needs to explain to McGraw-Hill that publishing a book without first proofreading it is an insult to readers and a blot on the corporate escutcheon. In conclusion, e-learning games are a dynamic field that is wide open to creative initiative and promises major long-term benefits. Prensky's book is a rich, insightful guide and makes absorbing reading.

The book states a clear and persuasive case for digital game-based learning and its ability to adapt to the varying needs of today's workforce. I highly recommend this book for anyone seeking an effective approach to learner-centered training.
This book goes beyond providing a great survey of some of the most exciting learning games created. The author does a great job comparing the cultural differences between the "Games Generation" and its predecessors.
I believe this would be a great book for managers working with employees from the Games Generation, especially those that question the generation's work ethic. Members of the Games Generation don't need work that is fun all the time, but it does need to be captivating. Getting stuck in a game isn't fun, but it is incredibly engaging. This information may improve working relationships and bridge some generational gaps.