First of all this is just my opinion, and sorry for my English.
I agree with Scott, that just looking at a DOE sample and running a DOE is not just that simple, or at least not if you want to obtain good results and not black magic.
Some question:
What is your problem ?
How much statistical knowledge ?
How much plastic injection manufacturing knowledge ?
I have seen several times when somebody was forcing a DOE, without processing knowledge, and also others running a DOE without any statistical knowledge. The third situation is when you put together a good DOE, but you miss to explain, give adequate information and supervision to those who will run the experiment, and will fail due to a small think not clear for all.
I these situations you may draw wrong conclusion. On the other hand running a DOE is very expensive and time consumung for shop floor.
You can find some examples, running a Google search, and here are a few, after I have run a search:
http://www.geocities.com/dfloyd2292/pap2.html
http://www.statease.com/pubs/doe-keys.pdf
http://www.public.iastate.edu/~vardeman/stat531/handnotes/doddstory.pdf
Good luck.
György