Completely Randomized Design with Subsampling – Example
These data come from an experiment that was conducted as a completely randomized design with subsampling: there were 4 treatments, and 4 plots for each treatment. Within each plot 3 measurements (subsamples) were taken. The treatments are different forms of irrigation; the subsamples correspond to 3 small amounts of soil that are measured in each plot — the measurement Y correspond to a measure of soil moisture. The data are tabulated below. For each treatment and plot combination, the 3 values represent the 3 subsamples.