INVESTIGATION OF DUAL-BALANCED CROSSOVER DESIGNS FOR 2 TREATMENTS

An overview of properties of two-period two-treatment crossover designs without and with baseline measurements is presented. In many situations, especially in biomedical experiments, investigators are often limited to the use of two-period designs. For two-period crossover experiments, the optimality of a design is largely affected by the presence or absence of carry-over effects. In this paper, a nearly optimal four-sequence two-period crossover design that is least affected by the uncertainty of carry-over effects is discussed.