Optimal minimal balanced crossover designs in first and second carryover effects

Abstract

Crossover designs balanced in first and second order carryover effects discussed in literature often require large number of experimental units. In practice experiments demand fewer experimental units where minimal balanced crossover designs are desirable. In this paper, a simple method of construction of minimal balanced crossover designs through a new form of terrace is presented. A new form of terrace constructs one series of crossover designs having less number of periods than the number of treatments and two series of crossover designs having more periods than the treatments. Crossover designs possess good efficiency of separability and optimal for the estimation of treatment effects.