|Design and Analysis
of Experiments (DOE)
Title: How to Analyze a Split-Plot Experiment
Authors: Scott M. Kowalski, Kevin Potcner
Publication Source: Quality Progress, Vol. 37, No. 12, December 2004, pp. 67-74
Quality improvement projects frequently require experimentation on a process. In many real experimental situations, a restriction is placed on the randomization of runs that affects the statistical analysis. A split-plot experimental method involving two error variances is a solution. The proper analysis of this type of design structure is discussed.