Abstract:
|
In industrial experiments, some factors are often harder to change than other factors. This situation often results in the experiment being run as a split-plot where the easy-to-change factors are changed more often than the hard-to-change factors. However, the scientist performing the experiment may not recognize the importance of incorporating the experimental protocol into the analysis and thus may incorrectly analyze it as a completely randomized design. This presentation offers a case study involving hydraulic fluid as a hard-to-change factor and compares the incorrect completely randomized analysis with the correct split-plot analysis.
|