Screening Designs are experimental designs optimized for identifying the vital few significant factors from a large number of potential factors — using a minimal number of runs to determine which of many candidate process variables actually affect the response, before investing in detailed optimization.

Key Screening Designs

• Fractional Factorials: $2^{k-p}$ designs that test $k$ factors in $2^{k-p}$ runs using aliases.
• Plackett-Burman: Economical 2-level designs in $N = 4n$ runs for up to $N-1$ factors.
• Definitive Screening: 3-level designs that can detect curvature and 2-factor interactions.
• Supersaturated: More factors than runs — for initial rough screening only.

Why It Matters

• Factor Reduction: Screening reduces 20-50 candidate factors to the 4-8 that truly matter.
• Efficiency: 12-run Plackett-Burman can screen 11 factors — far fewer than the 2048 runs for a full $2^{11}$ design.
• First Step: Screening is the essential first stage of any systematic process optimization.

Screening Designs are finding the vital few from the trivial many — efficiently identifying which process parameters truly drive quality from a large candidate list.