A researcher studied historical (relict) stands of longleaf pine in a South Carolina County. In each of 4 stands, they randomly selected two different 200-meter diameter plots and measured soil fertility for each plot.
In addition, the researcher randomly selected 20 200-meter diameter plots throughout the county (excluding the historical longleaf stands). Soil fertility for these sites was measured as well. The researcher wants to confirm whether relict longleaf pine stands are appropriate candidates for restoration (just because these stands lasted longer than others, they may not be the most appropriate candidates for restoration, from the standpoint of soil conditions).
To carry out the analysis, the researcher randomly grouped the 20 non-relict plots into 10 "stands" of size two. He then ran a nested model with Stand Type as the non-randomized treatment factor and Stand(Stand Type) as a random effect.
Comment on the researcher's design. Why might the researcher have grouped plots? Was is appropriate to do so? Why or why not? Propose an alternative analysis method and discuss its advantages/shortcomings.