A two-stage strategy is used to fit group orthogonal supersaturated designs. The first stage identifies which groups contain active effects.
• Obtain an estimate of σ from the fake factors and then discard the fake factors from further analyses.
• Test for significance of each group of factors.
The second stage considers significance of individual factors in active groups. This stage uses a guided subsets procedure. See Jones and Nachtsheim (2016). For complete statistical details, see Jones, et al. (2019).