ASA Connect

I would stick with Montgomery. A lot of my classmates have a math/stats only background and the examples in his book force those students to realize where the data comes from and issues that tend to happen in real world data. I'm trying to start a stat consulting group on my campus to get my classmates actively involved in designing the experiments too.

Box, Hunter and Hunter is good too, but can be a little above an intro to DOE class.

A good supplemental book would be Optimal Design of Experiments: A Case Study Approach by Goos and Jones. This book gets into where the levels for a design come from and why certain factors were chosen. It's actually a real fun and easy read too.   

Some other textbooks spend too much time on the math and how your software calculates the coefficients and not enough time showing why DOE is useful. They don't bother to show you how to use software either. I think that makes them fairly useless. (In the real world, we use software to calculate everything. Hand calcs take to long and we usually get the wrong answers. Plus, hand calcs will turn students off math/stats faster than anything.) 

I purchased several "Design of Experiments" books written by non-statisticians/industrial engineers, mostly biologists and chemists. Some of the authors say you can't change more than one thing at a time during an experiment! They also tend to get their own problems wrong. (Meaning if a student answered the questions the same way these authors did, you'd fail them for not transforming a response variable, using R^2 as the only measure of model adequacy, disregarding R^2(adj), residual plots, lack of fit, etc.) The books are a good source of practice problems and having students deal with mistakes scientists make with their data and analyses, constructive criticism of what they would do differently, and how to do things better.   

------------------------------
Andrew Ekstrom
------------------------------