Central composite design with three stars

Abstract

The central composite designs, here considered, present factorial points at levels ± 1, three stars at levels ± α, ± γα, ± ґ α, and central points. In fertilizer research, the responses may vary from year to year due to climatic and soil variability. If a proper range of dosages is utilized, in normal or good years the obtained response surfaces increase up to the high dosages; in unfavorable years, they may present a plateau area or even decrescent response beyond a certain level. The proposed designs fit these peculiarities well because of their structure and number of levels; the elimination of points in the decrescent or in the plateau area allows the determination of unbiased estimates for the model coefficients, thus furnishing more realistic economical decisions. Orthogonal, completely randomized, and orthogonal or quasi-orthogonal, orthogonally blocked designs are presented in three tables.