3-LEVEL SCREENING DESIGNS FOR THE OPTIMIZATION OR THE RUGGEDNESS TESTING OF ANALYTICAL PROCEDURES

Several Plackett-Burman designs have been proposed as screening designs for the optimisation or the ruggedness testing of a test procedure. One of these, a three-level design, described in the literature can lead to wrong conclusions about the influence of the different factors due to the construction of the design. When three levels of the variables that may have an influence on the test procedure require examination, one should use a reflected two-level design or a well balanced three-level design. A proposal for the latter is given.