Randomization assigns run order, or which units get which treatment, by chance. It protects against lurking variables such as tool warm-up, drifting chemistry or operator fatigue, which would otherwise line up with factor settings and fake an effect. When a factor is hard to change, such as furnace temperature, full randomization is often restricted, giving a split-plot design that needs a different error term in the analysis.
Don't run all the low-temp wafers first. Randomize the order, or the chamber seasoning shows up as a temperature effect.
Also heard as
- random run order
- randomized
Related terms
Block
Group of experimental runs made under similar conditions, used to remove a known nuisance source such as day, batch or tool.
Confounding
Mixing of two effects so that the data cannot separate them, whether by design or through an uncontrolled variable.
DOE
Design of experiments: planning a set of runs that vary several factors at once so their effects can be separated.
Replicate
Independent rerun of the same factor settings from scratch, which gives a true estimate of experimental error.