- How it Works
- Applications Overview
- Bioprocess Development
- Molecular Biology
- There are lots of books on DOE out there, aimed at a variety of audiences
- In this blog, we pick seven of our favorites and tell you a bit about why we love them and who they would be most suitable for
- Want something to get into straight away? Don’t forget our ebook, A Biologist’s Guide To Design of Experiments, or our DOE Masterclass webinar series.
Design of experiments (DOE) is a rich subject with a deep literature. Drawing on results and ideas from statistics, mathematics, and optimization, there’s a lot to learn—no matter how much you’ve already read about it. While the primary literature is vast and technical, there’s an enormous practical interest in the topic and a pressing need to get it into the hands of working experimenters. This has led to a huge variety of excellent textbooks being written, many of which are aimed at audiences with varying degrees of theoretical background. It can be hard to know where to start.
We’ve amassed quite the bookshelf on DOE over the years, both physical and virtual. So here’s a selection of books we think you should know about. They’ll be ideal if they’re your first (or second) book on the topic.
DOE Simplified (Anderson & Whitcomb)
DOE Simplified is the most-requested book by Synthace starters looking to get a handle on DOE. It’s also the book that got our CSO Markus into DOE in the first place! I hadn’t actually read it before I decided to write about DOE books so it was really interesting to do that, and find out why.
DOE Simplified was written by two members of Stat-Ease, who make Design Expert. Given the focus and success that Design Expert has always had on making DOE accessible, it makes a lot of sense that they would produce such an accessible book. Of course, the book is linked with the software, and readers are invited to download a free trial to play along at home.
That said, I was a little surprised that this book isn’t as introductory as expected: it’s got a great deal of statistical detail in it. The ‘Simplified’ part comes from trying to present all of the relevant statistical background in an accessible way. It also goes through things “properly,” building up the basic statistical foundations before moving on to the more DOE-specific stuff.
Overall, I think this works really well. The style is very readable and accessible and everything is backed up with specific examples, both as tables showing calculations and, importantly, as real experimental examples.
My biggest criticism is that it buries the lead. The initial flowchart guide to the book, for example, assumes a lot of DOE knowledge and focuses on details. While the book does have a very useful overview of the process as a whole, it’s all the way back in chapter ten. While building "from the ground up" like this is quite normal for textbooks of this kind (most of the books on this list do the same thing) for me it would have been even better to give more of an outline before getting into the details.
Still, it’s a very accessible book that has a lot to interest those new to DOE but which is nonetheless likely to interest more experienced practitioners as well.
The Rest of the “...Simplified” Series (Anderson & Whitcomb)
Following the success of DOE Simplified there are two further books in the series: RSM Simplified and Fomulation Simplified (with Martin Bezner).
RSM Simplified is, for me, a better book to start with than DOE simplified. It begins with a good overview of the RSM process, which is more generally useful than the statistical specifics of DOE Simplified. For my way of thinking, it’s better to understand the context of each type of experiment before you try to understand the details involved in doing it.
Overall, RSM Simplified is an excellent read. It definitely gets more technical than DOE Simplified does in later chapters, but does a better job at giving a high-level overview of what’s going on, which is great for beginners. For more seasoned readers their discussion of the ins and outs of rotatability is worth the book on its own in my opinion.
Formulation Simplified is a different beast altogether: definitely non-introductory, this is a deep dive into mixture designs, a very important special case beyond the more general-purpose applications in the other two books. I must confess I still have to read this properly so I can’t comment on details but my impression is it’s absolutely in line with the others in the series in terms of presenting technical matters accessibly. However, unlike the other two, it’s most definitely special interest only. If you need to read this, you’ll know.
NIST Engineering Statistics Handbook (free!)
This is not strictly a DOE book because it covers a variety of statistical topics. But, if you’re comfortable with stats and not put off by equations this is a great first book to read. It gets right to the point without doing the usual math-book thing of having one accessible page followed by a thousand pages of unexplained symbols. Chapters four (process modeling) and five (process improvement) are particularly useful for their introductory information on DOE and RSM.
A few of us here have read this either as a first or second book and usually refer it to others. It works well as a first book because it does a lot to provide context around where the different methods come from and how they fit together before getting into the details. It also gives some useful examples. The downside is it *is* an engineering book so all of the examples cover things like manufacturing and materials science, but they’re pretty simple and well-explained so they still illustrate the ideas and methods well.
Even if you’re not keen on equations it’s definitely worth a quick perusal of the explanatory sections of the chapters, there’s really no better concise explanation of the underlying ideas of DOE that I know of. Also, it’s free! What’s not to like?
Statistics For Experimenters (Box, Hunter & Hunter)
The classic book for DOE in many ways, and rightly so. This is a very readable but still comprehensive and rigorous book that covers a huge amount of ground from statistical fundamentals through to more advanced and specialized topics like process control, designing for robustness, and evolutionary operation (which Box wrote another great book all about).
Unashamedly mathematical, this is definitely a great book to read for anyone at almost any level - the material is generally well explained with plenty of exercises, examples, and experiments to try at home, but includes lots of useful tidbits and rules of thumb as well as the specialized topics, making it still of interest to experienced DOE practitioners.
An excellent foundational text, probably an alternative to DOE / RSM Simplified for an introduction but starting to show its age a little: it’s firmly rooted in the more classical DOE methods and missing treatments of optimal designs, for example.
One useful feature is their chapter on nonlinear designs, which is great. Another is the introductory chapters on the advanced topics I mentioned before. Overall, I’d say Chapter one (“Catalyzing the growth of knowledge”) is well worth a read for anyone. Easy enough to find online in PDF form so you don’t need to pay the excruciating second-hand prices it fetches in hardcopy.
Optimal Design Of Experiments: A Case Study Approach (Goos & Jones)
Until the more recent publication of Design of Experiments: A Modern Approach this was by far the most accessible and general introduction to using optimal designs. Goos and Jones aimed to produce a book that would cater to both beginners and experts, a tough brief as they admit, but to some extent successful.
My guess is that they probably overestimate the statistical interest and experience of new practitioners somewhat. But as an introduction to people with the requisite statistics background, it is an absolutely excellent summary with more than enough to get to grips with the subject.
The unifying theme of case studies is a good expository device and while the mocked-up ‘consultancy’ bits don’t entirely ring true as dialogue to my ‘inner ear’, the various ideas and questions they discuss are very stimulating, and the format works well for what could be a dry subject. Each case study is carefully chosen to highlight a specific and often misunderstood fact about design and analysis, and there are so many interesting and surprising details I found it really hard to put down.
Design of Experiments: A Modern Approach (Jones & Montgomery)
I’ve not read this one as thoroughly as I’d like, so I can’t necessarily critique some of the details as well as with some of the others, but it’s well worth knowing about because I think the approach they authors have taken is spot on.
Essentially, all of the books above give a lot of emphasis to ‘classical’ DOE designs and methods. They take you through everything you need to know to actually sit down and do your DOEs with a pen, some paper, and the appropriate reference materials such as tables of designs.
This is a fantastic approach if you are traveling back in time to 1980 to do your DOEs but, here in 2022, we all have access to more than enough CPU power and great software to not really need to do things that way. Therefore a book that takes this basic fact as a starting point is incredibly welcome.
This has two important consequences: first, they don’t need to waste time on the details of how to calculate everything manually. Second, they use optimal designs as a jumping-off point. This is very welcome, as none of the more accessible textbooks really give you much on how to approach using these. In our experience, optimal designs are often the only real choice for the experiments our colleagues want to do so I anticipate this will become an important book for users of DOE in biosciences.
Although this is intended as an introductory textbook it’s probably better read after DOE Made Simple or RSM Made Simple if you’re coming at it without a lot of statistical background; it goes into the details pretty quickly. However, the first chapter does score very highly as an overview of the main concerns of experimental design and is definitely a great introduction.
You’re gonna need a bigger bookshelf
We’ve been through a handful of books either entirely or mostly about design of experiments here, including some of the best current resources alongside some classic choices. But of course we’ve barely scratched the surface! There are plenty of others out there, and we can only apologize for our omissions (especially to you, Wu & Hamada). Hopefully if there’s interest we’ll follow up with some others in future.
If you’re looking for something accessible that you can download immediately, we’d also recommend our ebook: A Biologist’s Guide To Design of Experiments. We wrote it with life science experimenters in mind. You can also watch the entirety of our DOE Masterclass webinar series where we walk through the fundamentals of what DOE is, and how it can be applied in biological experiments.
Director of Scientific Software at Synthace
Other posts you might be interested inView All Posts
International Women’s Day 2022: Break the Bias, Break Barriers
Design of Experiments (DOE)
Here’s Why You Should Stop Doing Science Like You Were Taught in High School: Part 3
Lab of the Future
How to Work in Your Lab… Without Being in Your Lab