Screening designs constructed using the Custom Design platform are often equivalent to the classical designs provided in the Screening Design platform. However, Custom Design can construct designs for cases where classical screening designs are not available.
In the early stages of studying a process, you identify a list of factors that potentially affect your response or responses. You are interested in identifying the active factors, that is, the factors that actually do affect your response or responses. A screening design helps you determine which factors are likely to be active. Once the active factors are identified, you can construct more sophisticated designs, such as response surface designs, to model interactions and curvature.
The Custom Design platform constructs screening designs using either the D-optimality or Bayesian D-optimality criterion. The D-optimality criterion minimizes the determinant of the covariance matrix of the model coefficient estimates. It follows that D-optimality focuses on precise estimates of the effects. See “Optimality Criteria”.
The following examples illustrate constructing screening designs using the Custom Design platform.
• Design That Estimates Main Effects Only
• Design That Estimates All Two-Factor Interactions
• Design That Avoids Aliasing of Main Effects and Two-Factor Interactions