This document discusses generalized linear mixed models (GLMMs) and their use in analyzing experimental data with more than one random source of variation, such as blocking or split plots. It provides an example of using a GLMM to analyze a split plot experiment with a proportional response variable. The GLMM allows modeling of non-normal data like counts or proportions while accounting for random effects. The document demonstrates how to set up the linear model for this type of split plot proportional data and analyze it using JMP's mixed modeling tools.