Design of experiments (DOE)

planned trials that change several factors at once

A structured way to find out which process settings change a result and by how much. Instead of changing one factor at a time, a designed experiment changes several factors together in a planned pattern, so that every run contributes to the estimate of every effect and interactions between factors become visible.

Typical steps: define the response, choose factors and levels, decide on replicates, randomise the run order, run the trials, analyse the effects and confirm the best setting.

Example

Instead of 3 separate trials for temperature, pressure and cooling time, one 2³ design with 8 runs tests all three and their interactions.

Common mistake

Treating DOE as a statistics exercise only. Most failed experiments fail on planning: levels too close together, a poor gauge, or settings that drift during the trial.

Calculate it

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

Related terms

• Factorial design (2^k): every combination of the factor levels
• Main effect: average change when one factor goes from low to high
• Interaction effect: the effect of one factor depends on another
• Randomisation: running the trials in random order