A1188
Title: Two new classes of mixed-level screening designs inspired by definitive screening designs
Authors: Christopher Nachtsheim - University of Minnesota (United States) [presenting]
Bradley Jones - Adsurgo LLC (United States)
Ryan Lekivetz - SAS Institute Inc. (United States)
Dibyen Majumdar - University of Illinois Chicago (United States)
Abstract: Two families of orthogonal, mixed, two- and three-level screening designs are introduced. Both classes of designs leverage the structure of definitive screening designs. The first family is a class of orthogonal, mixed-level screening designs in multiples of eight runs that provide substantial bias protection of the main effects estimates due to active two-factor interactions. The second is a family of saturated designs containing $m$ two-level continuous factors and $m-1$ two-level categorical or continuous factors in $n = 2m$ runs, where m is greater than or equal to four. A key advantage is that these designs are available for any even $n \ge 8$. For $n$ a multiple of four, the designs are shown to be as effective as the best two-level alternative, namely Hadamard-matrix-based designs. It is shown that the designs typically have power near one for identifying up to $m$ active main effects when the signal-to-noise ratio is greater than 1.5.
