Randomisation

running the trials in random order

Running the experiment in a random order spreads anything that changes during the trial, such as warm-up, tool wear, a new material lot or a shift change, over all factor settings. Without it, such drift can look exactly like a factor effect.

If a factor is very hard to change, the runs can be grouped by that factor (a split-plot design), but the analysis must then account for it.

Calculate it

DOE: 2-Level Factorial Design: 2^k factorial: run plan, effects, Pareto, interaction plots and best settings.

Related terms

• Replicate: a complete repeat of the design
• Design of experiments (DOE): planned trials that change several factors at once