Definition
Design of experiments (DOE) is a systematic way to find out, with a minimum of runs, which factors (temperature, pressure, material) affect a result and how to set them. Instead of changing one factor at a time it changes all of them together according to a plan, so it also reveals interactions between factors. Evaluation is by ANOVA and regression; response surface methodology then locates the optimum.
Quality control & SPC
The basis is two-level factorial designs 2ᵏ and their fractional variants for screening many factors, then central composite and Box–Behnken designs for modelling curvature, mixture designs, and Taguchi orthogonal arrays for robust design.
A good experiment randomizes the run order, replicates to estimate error and blocks known nuisance influences (batch, shift). The outputs are a Pareto chart of effects, interaction plots and response surfaces that show where the process performs best and is least sensitive.
In Statistica
Statistics → Industrial Statistics & Six Sigma → Experimental Design (DOE) generates factorial, fractional, central composite, Box–Behnken, Taguchi and mixture designs, evaluates them by ANOVA with a Pareto chart of effects, interaction plots, response surfaces and contour plots, and offers desirability optimization for several responses at once.
Related terms
- ANOVA (analysis of variance)ANOVA (analysis of variance) tests whether the means of three or more groups differ.
- Linear regressionLinear regression describes how a dependent variable (yield, price, consumption) depends on one or more explanatory…
- Six SigmaSix Sigma is a data- and statistics-driven methodology for process improvement.
- Process capability (Cp, Cpk)Capability indices compare process variability with the customer's tolerance limits.
Knowledgebase guides
FAQ
- How many runs do I need?
- A full factorial design for k two-level factors has 2ᵏ runs; for 5 or more factors use a fractional design (e.g. 16 runs for 7 factors) for screening and optimize only the selected ones.
- Why is changing one factor at a time not enough?
- Because you miss interactions — the optimal temperature may depend on pressure. DOE reveals interactions with fewer runs than sequential trials.
Try it on your own data
Statistica free for 30 days
Full version, no credit card. Or get a pricing estimate in a minute.
Quality control & SPC
Updated: September 2026.