Professor Douglas Montgomery is a Regents’ Professor of Engineering at Arizona State University, who is known for his research on Engineering statistics, including design and analysis of experiments, statistical methods for process monitoring and optimization, and the analysis of time-oriented data.
He is also the Editor of Quality and Reliability Engineering International and the recipient of numerous awards, including the American Society for Quality Control’s Shewhart Medal. He also works as a consultant and is the author of many books, such as the best-selling Design and Analysis of Experiments.
Statistics Views talks to Professor Montgomery about his career in engineering and statistics.
1. With an educational background at Virginia Polytechnic Institute, when and how did you first become aware of statistics as a discipline and what led you to pursue a career within the subject area?
As an undergraduate, engineering students were required to take a course called Engineering Statistics, which was a junior-level course for me. This course was taught by a faculty member named Ray Myers. We are now co-authors of a couple of Wiley books for the Probability and Statistics series. I just thought that the subject was absolutely fascinating because I had done a couple of summer jobs with a division of union carbine as an engineering student and I saw right away some of the issues we had covered on the course could have been used in some of the activities I was involved in with carbine.
So I took a subsequent elective course in my senior year and became even more interested, so when I ended up in graduate school, it seemed natural for me to continue taking courses in statistics. I stayed in engineering and took an MS and PhD in engineering, however I took lots of courses in statistics, including from Ray Myers who was a member of my dissertation committee. I have been friends with Ray ever since and that friendship goes back to 1962.
2. In terms of your research, what are you currently focussing on? What are your main objectives and what do you hope to achieve?
I’m mainly working on experimental design problems and I have a research project with the Department of Defence to develop methods and techniques in what the military calls, ‘Operational Testing’. This is where you take a device/product/system out into the field and you test it with user troops to see how it performs and what its capabilities are. Obviously this is very expensive experimental work, so finding efficient ways to do that is a priority. During the last few years, the Department of Defence has placed a large emphasis on using statistically designed experiments in order to do this but a lot of the problems are fairly unique. They are complex and involve lots of factors, so developing methods that can be used for these large experiments is an area that I’ve worked in for the last few years. I even have a couple of doctoral students who are still working on some aspects of this.
3. You have authored several publications for Wiley including Engineering Statistics, Statistical Quality Control and Design and Analysis of Experiments, many of which have produced multiple editions and the sixth edition of Applied Statistics & Probability for Engineers has also just been published. Is there a particular article or book that you are most proud of?
I think it would be Design and Analysis of Experiments. It really changed the way people teach experimental design, particularly to students studying engineering and science who desperately need it! At the time I wrote the book, back in the 1970s, there really wasn’t a book on experimental design that was orientated towards engineers and physical/chemical scientists. The books were very agriculturally focused and with a roomful of engineers, you don’t want to walk in and start off with an example, “So, suppose you have 4 sheep…” they don’t relate to that! So I started finding examples of where this could be used in engineering problems and bringing them into class, so finally I decided, ‘What the heck? I should write a book on this’ and so I did. And I’m proud of its success, it’s now in its 8th edition and I think it is the most widely used book on the subject or if not, it’s very close.
4. What is it about the subject of engineering statistics that continues to fascinate you to this day?
It’s the interplay between the engineering science and problem-solving because that’s basically what engineers and many scientists are. Engineers are basically problem-solvers and almost every problem that you run into that is meaningful has some aspects of having to deal with data. You need to know how to do that in an efficient and effective way. You can’t really do that without knowing something about statistics so my whole career has been focused on trying to get these techniques more widely known in the engineering and science community. It continues to be fascinating to me that new problems crop up that we can use and I find them fun to address.
5. What will be your next book-length undertaking?
I am still doing some revisions on a book on time series and forecasting that I am a co-author of for Wiley. I am also working on a new book with a colleague at SAS in the JMP division, named Brad Jones on optimal designs. It’s a new way to teach the basic principles of design from an optimal design viewpoint. We have been working on this book for a couple of years but sadly both of us have been incredibly busy so it’s been subject to some delay but we will finish it. I am really enthusiastic about the book.
During the last few years, the Department of Defence has placed a large emphasis on using statistically designed experiments in order to do this but a lot of the problems are fairly unique. They are complex and involve lots of factors, so developing methods that can be used for these large experiments is an area that I’ve worked in for the last few years.
6. You are currently Regents’ Professor of Engineering and ASU Foundation Professor of Engineering at Arizona State University. Do you have any advice for students considering a university degree in engineering and statistics?
If you are an undergraduate student in engineering or any science, you need to be sure to take at least one course in statistics because it will be useful to you throughout your career. If you are a graduate school and involved in research, I would definitely recommend a course in experimental design as so much of your research work is experimental in nature. I teach an introductory course on experimental design every year and this year, I taught it both in the fall and the spring semesters and I had about 150 graduate students in each class. They were from all over the engineering disciplines, as well as chemistry and biology, so the more of that you can learn, the better.
We have an online MENG program here in Quality, Reliability and Statistical Engineering with a huge number of students pursuing that degree. I think we have somewhere between 70-100 students for this course, most of whom are already working in the industry and are studying part-time for this MENG. There is a big interest in the field, so I would recommend taking these courses because they will be tremendously useful to you.
7. Over the years, how has your teaching, consulting, and research motivated and influenced each other? Do you continue to get research ideas and incorporate your ideas into your teaching? Where do you get inspiration for your research projects and books?
Yes, definitely. My courses evolve on a regular basis and they are driven by the research that I do. A lot of the research is driven by problems in the industrial and business world, some of which comes through consulting activities and as I said before, I have this funded research project from the Department of Defence where I am focusing on the specific problems they have that are hard to solve. There has always been a nice interplay between consulting clients, research and teaching and all of that somehow integrates into the textbooks I write. I work really hard to keep the textbooks up-to-date methodologically and also to keep them up-to-date in terms of the types of applications. That is why there are so many editions!
8. What is the best book on engineering statistics that you have ever read?
The book that I learned from as an undergraduate student was Engineering Statistics by Albert H. Bowker and Gerald J. Lieberman, an old book published by Prentice-Hall and it was first published in the late 1950s or early 1960s. That was the book I studied from and had a real influence on me because it was well-written, clear, and concise, had excellent real-world examples – everything a good textbook should be and in a way, it became a role model as to how to write a textbook. It was clear that Lieberman had a lot of practical experience because you could see it in the way the book was written. I think it was certainly the best book of its era.
9. What do you think the most important recent developments in the field have been? What do you think will be the most exciting and productive areas of research in engineering during the next few years?
One of them would be the greatly increased use of computer models in engineering – engineers use computer models for all kinds of design activities and to study processes in systems that they cannot observe easily. For example, there are computer software packages that can be used to assist engineers to design electrical circuits and in my specific area of industrial engineering, people are using computer models to study logistics operations, supply chains, service systems including health care delivery systems. The question is how to use these models in a sensible and intelligent way – you run experiments and now we’re back in the world I operate in, the design of experiments – and there is a growing field within experimental design on experiments for computer models. This is a fairly recent development in our field where statistics and engineering are coming together in a very proactive way. I have done some research in the area myself and it is interesting as there are a lot of new problems that we had not thought about before.
10. What do you see as the greatest challenges facing the profession of statistics in the coming years?
How to deal with Big Data, certainly! Big Data is a huge challenge for all of us – how do we make use of this intelligently and what sort of techniques and methods do we need to develop in order to tackle Big Data?
11. How do you manage to juggle editing Quality and Reliability Engineering International and your many other commitments?
I try to spend about a day every two weeks catching up with the journal. I have to allocate this time from dawn to dusk and work on everything as best I can. Sometimes I am successful, sometimes I am not but I do the best I can!
12. Are there any hot areas that you’d like to see the journal publish in?
I would like to see us have more papers dealing with Big Data – data mining, for instance. A lot of these now appear in computer science journals and a lot of them could certainly be published in statistics journals. I would also like to see more papers on computer experiments, particularly applications. More novel case studies dealing with computer models would be great – we get a modest number of case study papers submitted but very few of the ones I process rise to the level of case study that I want to see in the journal. I don’t want to see case studies on routine applications that readers won’t really benefit from. I would like to see case studies on novel applications of existing techniques or their new problems where new methods have to be used in a clever way in order to be able to solve.
For Quality and Reliability Engineering International, I would like to see us have more papers dealing with Big Data – data mining, for instance. A lot of these now appear in computer science journals and a lot of them could certainly be published in statistics journals. I would also like to see more papers on computer experiments, particularly applications.
13. QREI has a couple of special issues on Nonparametric Statistical Process Control Charts and Data Mining in the pipeline but are there any other topics that you’d like to cover?
We do 2-3 special issues a year which I think promotes interest in the journal. We very often have a special issue tied to the ENBIS conference and that is always a good issue. Any time someone comes to us with an idea for a special issue, we are more than willing to listen to the ideas.
14. Who should read the Journal and why?
I think we see the journal aimed at both practitioners and researchers which is a balance that is hard to achieve. The research papers have to be written in way that they are very succinct. We want a more informal writing style and researchers need to make their methods accessible to people in the field. If the only way you can figure something out is working through 3-5 theorems, then the practitioner will lose interest quickly and we don’t want that!
15. Are there people or events that have been influential in your career?
I have to back to my days at Virginia Tech and some of the engineering faculty I studied with were very influential, such as Richard Leavenworth, who was a co-author on Eugene Grant’s book on Quality Control and a wonderful teacher. Dr Roger Smith, who taught the undergraduate quality control course, was a great teacher and mentor to me and highly influential. Prab Ghari who worked on operations research at Virginia Tech and of course, Ray Myers – these four really taught me a lot about our profession. When I left Virginia Tech and worked at Georgia, I worked with Lynwood Johnson and Bill Hines and they were also influential – in fact, I have authored books with both of them! That is probably just a short list as there are many people over the years that I have learnt a lot from, for example, my colleague, Brad Jones and some of my colleagues here at ASU have been very influential on my career, such as Burt Keats who was the reason I came to ASU. I owe these folks a lot of credit for anything that I have accomplished.
Copyright: Image copyright of Arizona State University