In experimental design, what defines a split-plot design?

• A. Subplots are independent of whole plots.
• B. It is identical to a randomized complete block design.
• C. It is a factorial design with full replication.
• D. Whole plots of replicates contain subplots of replicates.

Answer: (C)

The main idea is that a split-plot design combines two levels of experimental unit: whole plots and subplots, with one factor applied to the whole plots and another factor applied within those plots. This creates two sources of randomization and two error terms, but the overall layout still reflects a factorial structure that is replicated across blocks.

The statement that best fits is that it is a factorial design with full replication. In a split-plot, you have a factorial arrangement of the factors, and this entire factorial is replicated across whole-plot blocks. Within each block, the subplot factor is applied to subplots, so treatments are organized in a way that preserves a complete factorial pattern across replicates, even though the randomization and error terms differ between whole plots and subplots.

Why the other descriptions don’t fit as cleanly: the split-plot’s subplots aren’t independent of the whole-plot treatment, because they share the same whole-plot factor; it’s not identical to a randomized complete block design due to the two-stage randomization and distinct error terms; and a phrasing like whole plots containing subplots of replicates is too vague and misses the explicit factorial arrangement and replication structure that define split-plot designs.