Previous Page

|

Next Page

Chapter Topics

In many experiments, treatments are applied to different sizes of experimental material. This occurs, for example, when some treatment levels are hard to change or when different batch preparation methods are studied. Some treatments are randomized on large units of material, called whole plots, while others are randomized on smaller units, called split plots or subplots. Split-plot designs are used to carry out such experiments.

Split-plot designs collect information from both whole plots and subplots, with corresponding nominal levels of noise. This is modeled by treating whole-plot errors as random effects. Therefore, ADX uses a mixed-models formulation to analyze these experiments.

Covered in This Chapter