Group orthogonal supersaturated designs are a special case of screening designs. They are appropriate in early stage work when the number of factors to be investigated is larger than the number of feasible runs. A supersaturated design is a design with fewer observations than model parameters. That is, the design cannot estimate all main effects simultaneously. This makes detection of significant effects difficult. A group orthogonal supersaturated design is a special class of two-level supersaturated designs with properties that are desirable for model selection.