Full Factorial Experiments

Full factorial experiments are the gold standard for understanding how multiple factors influence a response. They test every possible combination of factor levels, providing the most complete and unbiased view of the process. In the Improve phase, full factorial designs are invaluable when you need to understand not only main effects but also interactions—situations where the effect of one factor depends on the level of another. 

The strength of full factorial designs lies in their completeness. By testing all combinations, you ensure that no effect is confounded with another. This allows you to estimate main effects, two‑way interactions, and, with additional levels, curvature. Full factorial designs provide the richest possible insight into the structure of the process. 

However, full factorial designs can be resource‑intensive. The number of runs grows exponentially with the number of factors. For example, a 2ᵏ design with five factors requires 32 runs. While this is manageable in many settings, larger designs may become impractical. This is why full factorials are often used when the number of factors is small or when the stakes are high enough to justify the investment. 

Full factorial designs also support clear, intuitive interpretation. Main effects plots show how each factor influences the response. Interaction plots reveal whether the effect of one factor depends on the level of another. These visual tools help you understand the process at a glance and communicate findings to stakeholders. 

In the Improve phase, full factorial experiments provide a comprehensive, unbiased foundation for improvement decisions. They help you understand the process deeply, identify key drivers, and design interventions that address the most influential factors.