Adding 3 to 5 center points to a two-level factorial lets the experimenter check whether the response is curved: if the center average differs significantly from the average of the corner runs, a linear model is not enough and the work moves to a response surface design. Repeated center points also supply a pure error estimate without replicating the whole design, and spreading them through the run order helps reveal drift.
Center points came in 8 percent above the corner average. There's curvature, we need axial points.
Also heard as
- centerpoint
- center run
Related terms
Factorial design
Experiment that runs every combination of the chosen factor levels, so each main effect and interaction can be estimated.
RSM
Response surface methodology: sequential experiments that fit curved models to find the settings that optimize a response.
Replicate
Independent rerun of the same factor settings from scratch, which gives a true estimate of experimental error.
DOE
Design of experiments: planning a set of runs that vary several factors at once so their effects can be separated.