Split plot experiments happen when it is convenient to run an experiment in groups of runs (called whole plots) where one or more factors stay constant within each group. Usually this is because these factors are difficult or expensive to change from run to run. JMP calls these factors Hard to change because this is usually how split plotting arises in industrial practice.
In a completely randomized design, any factor can change its setting from one run to the next. When certain factors are hard to change, the completely randomized design may require more changes in the settings of hard-to-change factors than desired.
If you know that a factor or two are difficult to change, then you can set the Changes setting of a factor from the default of Easy to Hard. Before making the design, you can set the number of whole plots you are willing to run.
For an example of creating a split plot design, see Creating a Design with Two Hard-to-Change Factors: Split Plot.
In the factors table there is a column called Changes. By default, changes are Easy for all factors. If, however, you click in the changes area for a factor, you can choose to make the factor Hard to change.
Note: If you enter a missing value in the Number of Whole Plots edit box, then JMP considers many different numbers of whole plots and chooses the number that maximizes the information about the coefficients in the model. It maximizes the determinant ofwhere V -1 is the inverse of the variance matrix of the responses. The matrix, V, is a function of how many whole plots there are, so changing the number of whole plots changes V, which can make a difference in the amount of information a design contains.