Overview of Screening Designs

Each continuous factor in a screening design is typically set at two levels. Classical designs can also accommodate categorical factors with three levels. When the number of possible factor level combinations results in an infeasible design size, response measurements are taken for only a fraction of these combinations.

The screening design platform provides screening designs for two-level continuous factors, categorical factors, and discrete continuous factors. These designs are of two types:

•	Classical designs: For situations where standard screening designs exist, you can choose from a list of these designs.

•

Main Effects Screening Designs: For situations where standard designs are not available, JMP generates designs that are either orthogonal or near orthogonal. Like some of the standard designs, these designs focus on estimating main effects in the presence of negligible interactions. See Near-Orthogonal Designs.

The emphasis on studying main effects early on in the experimentation process is supported by the Hierarchical Ordering Principle (Wu and Hamada, 2000). This principle contends that lower effects are more likely to be important than higher order effects. For this reason, screening designs generally assume that interactions are negligible. In cases where this assumption is not reasonable, screening designs assume that two-factor interactions are more important than three-factor interactions, and so on.

The efficiency of screening designs also depends on the critical assumption of effect sparsity. Effect sparsity reflects the fact that real-world processes usually have only a few driving factors. Other factors are relatively unimportant. To understand the importance of effect sparsity, you can contrast screening designs to full factorial designs:

•	Full factorial designs consist of all combinations of the levels of the factors. The number of runs is the product of the factor levels. For example, a full factorial experiment with three two-level factors and a four-level factor has 2 x 2 x 2 x 4 = 32 runs.

•	By contrast, screening designs require only a fraction of the runs in the full factorial design. The main effects of the four factors described above can be studied in an eight run main effects screening design.

This chapter shows you how to generate and analyze screening designs, pointing out their limitations in the process.