Rf. Gunst et Gc. Mcdonald, THE IMPORTANCE OF OUTCOME DYNAMICS, SIMPLE GEOMETRY AND PRAGMATIC STATISTICAL ARGUMENTS IN EXPOSING DEFICIENCIES OF EXPERIMENTAL-DESIGN STRATEGIES, The American statistician, 50(1), 1996, pp. 44-50
One-factor-at-a-time experiments and similar ''common sense'' design s
trategies continue to be prevalent in industrial experiments in spite
of the strong emphasis in statistics courses that these design strateg
ies should be avoided. In justifying the avoidance of such design stra
tegies, technical criteria such as design efficiency and confounding a
re usually stressed, However, these technical criteria routinely are o
verlooked by experimenters in industry when the desire for rapid feedb
ack on factor effects is given undue precedence, In this article the p
reference or rapid feedback over statistical considerations is discuss
ed relative to the compelling inherent deficiencies of these designs,
The practical difficulties associated with the dynamic nature of the d
esign construction, geometric imbalance of the designs, and serious mo
deling inadequacies that result om the use of these design strategies
are detailed.