Researchers at a plastics manufacturer want to increase the strength of a plastic. The researchers identify additive percentage, agitation rate, and processing time as the possible factors that affect strength. The temperature at which the plastic bakes also affects strength. To run a completely randomized 4-factor design requires that the researchers bake each combination of the within-batch factor levels individually at one of the two temperature settings. Because the process takes too long, the researchers decide to use a split-plot design. The researchers plan to bake all 8 combinations of additive percent, agitation rate, and processing time at one temperature, and then bake all 8 combinations at the second temperature. They replicate this process so that they use each temperature setting twice. This design results in 32 observations, run in 4 whole plots of 8 subplots each.
The first table provides a summary of the design.
Because the design is a full factorial, the alias table says that the terms are free from aliasing.
The Design Table shows the experimental conditions or settings for each of the factors for the design points. The first 8 runs of this split-plot experiment represent the first whole plot, and Factor A, which is a hard-to-change factor, is set at the high level. The first subplot run in the first whole plot has Factor B high, Factor C high, and Factor D low.
Minitab randomizes the design by default, so if you replicate this example your run order will not match the order in the example output.