Agreed on the Engineering Statistics Handbook, especially Section 5 on Process Improvement, which contains a decent (and free) treatment of Design of Experiments. A lot of the earlier sections are ok but not as sharply interesting, but their DoE stuff is quite good.