What is a Good Experiment Design?
A good experiment design gets you the answer you need at a price that you can afford to pay. A better experiment design gets you the answer you want for less than you were willing to pay. The best experiment design gets you the best answer at the least possible cost. Statisticians refer to designs in this last category as OPTIMAL and MINIMAL. Let’s start with the elements of a good design. First, the experiment needs to have a clearly defined objective. Without a clearly defined objective and some assurance that the results will be used, the experiment and thus the design are without purpose. This seems obvious, but more experiments go wrong at this point than at any other. Second, the experiment must consider ALL of the factors that may contribute to the outcomes of concern. Why would you leave any out? To make the design simpler and to save money. Experiments with fewer factors are cheaper (more factors mean more trials), and different kinds of factors can add complicating wrinkles t
