OFAT (One Factor at A Time)

OFAT = One Factor at A Time

Supporting References

• Hicks page 133

• Box page 312

• NIST (5.2.1.2)

DOE = Design of Experiments, aka Statistically Designed Experiment

• OFAT requires more effort and data, while providing less information than DOE.

• OFAT is not 'orthogonal' and usually does not allow the estimation of interactions between factors.

• OFAT extrapolates within the design space, while DOE interpolates.

• Using three or more levels in OFAT commonly leads some to believe that they can estimate higher-order effects (quadratic, cubic, etc.), but this is not an effective method. See Three Level Factorial Design Matrix for an improved approach. Also note Central Composite designs (CCD).

• OFAT does not provide an effective way to add to the Design Structure in subsequent designs.

• OFAT tends to ignore the many pitfalls of not using a randomized order to testing the various treatments.