Definitive Screening (a.k.a. DSD)
These designs allow all main effects to be estimated, clear of second-order effects. The second-order effects (two-factor interactions and quadratic terms) are aliased with each other. If second-order effects appear to be significant, proceed with caution.
Significant second-order effects reveal that the relationship between the factors and the response requires more than simple linear terms in the model. However, correct identification of the second-order effects requires that only a few factors truly contribute to the second-order and/or there is sufficient subject matter knowledge to make the decision .
Check the evaluation results before using these designs to make sure they can detect significant effects at the size needed to make a decision. As with any design created to minimize the runs, a definitive screening design runs the risk of not having enough power to detect small but important effects.