JMP for Mixed Models

Chapter 8Simulation and Power Analysis

We have left discussion of sample size and power to later in the book, but in reality these calculations should be done at the design phase of the experiment, long before analyzing data. If the experiment that you run is not going to have the power to detect the difference you are hoping to see, then all of the time and materials used in the experiment are potentially wasted. Power and sample size calculators are available in many software packages, including JMP, but most of these options do not take mixed model analyses into consideration in their calculations. When there is more than one source of variability, tests (and confidence intervals) for means and their differences often have complex combinations ...