Modern Experimental Design
Chapter One
Introduction
The statistical design of experiments plays a prominent role in experimentation. As
George Box has stated, to see how a system functions when you have interfered with
it, you have to interfere with it. That "interference" must be done in a systematic way
so that the data from the experiment produce meaningful information.
The design of an experiment should be influenced by (1) the objectives of the experiment,
(2) the extent to which sequential experimentation will be performed, if at all,
(3) the number of factors under investigation, (4) the possible presence of identifiable
and nonidentifiable extraneous factors, (5) the amount of money available for the experimentation,
and (6) the purported model for modeling the response variable. Inman,
Ledolter, Lenth, and Niemi (1992) stated, "Finally, it is impossible to overemphasize
the importance of using a statistical model that matches the experimental design that
was actually used." If we turn that statement around, we should use a design that
matches a tentative model, recognizing that we won't know t ... read full excerpt from Modern Experimental Design ebook
