On the optimality of the generalized room squares

Abstract

Using a model given by Archbold and Johnson (1958) it is shown that the generalised Room square design is universally optimal for the estimation of the treatment effects. Further, if the variety-block design in a generalised Room square design is a BIBD, this design is also universally optimal for the estimation of the varietal effects.

Keywords:

• sub-plot analysis
• main-plot analysis
• universal optimality
• balanced incomplete block design