Bingham and Sitter (1999) described a wood product experiment using split plot
structure. The experiment involves a two-stage process: the mixing stage and the
processing stage. The objective is to study the effects of eight factors, with two
levels each. These factors are divided into two groups: five factors in Stage 1 and
three factors in Stage 2. Due to process restrictions, the experiment was conducted
by first randomly preparing batches with different treatment combinations of the five
factors in Stage 1. Then each batch was divided into sub-batches. Each sub-batch
was then applied a treatment combination of the three factors in Stage 2 to further
facilitate the forming process.