Constructs randomized nested row-and-column type block designs with arbitrary depth of nesting for arbitrary factorial or fractional factorial treatment designs. The treatment model can be defined by a models.matrix formula which allows any feasible combination of quantitative or qualitative model terms. Any feasible design size can be defined and, where necessary, a D-optimal swapping routine will find the best fraction for the required design size. Blocks are nested hierarchically in successive strata and the blocks model in any particular stratum will comprise either a simple nested blocks design or a crossed row-and-column blocks design. Block sizes in any one stratum are either all equal or differ, at most, by one plot. The design outputs include a data frame showing the allocation of treatments to blocks, a table showing block levels, the fractional design efficiency, the achieved D-efficiency, the achieved A-efficiency (unstructured treatments only) and A-efficiency upper bounds, where available, for each stratum in the design. For designs with simple unstructured treatments, a plan layout showing the allocation of treatments to blocks or to rows and columns in the bottom stratum of the design is also given.
|Author||R. N. Edmondson|
|Date of publication||2017-01-08 15:42:21|
|Maintainer||Rodney Edmondson <email@example.com>|
|License||GPL (>= 2)|