Building Models for Complex DOEs
Donald W. McCormack, Technical Enablement Specialist, JMP
Any experiment designed in JMP will provide you with the model needed to analyze that experiment. This is particularly helpful in cases where random effects are used for blocking or hard-to-change factors. Experiments coming from alternative sources often prove challenging to set up in JMP. While full and fractional factorial designs are relatively easy to implement, strip-plot and split-plot designs can be much more difficult. And what of designs not available in JMP, such as crossover designs, Latin squares, or split-strip plots? In these situations it is not possible to create a similar model with the Custom Design platform to see how the Fit Model dialog should look. This presentation will focus on designs with one or more random effects and demonstrate how to set up the Fit Model dialog. Examples will also include designs with more than two levels of nesting.