Crossover Design

An experimental design in which treatments are administered in sequence to experimental units. Experimental units will receive the treatments in varying orders.

The main advantages of this design is that 1) each treatment is given to each experimental unit (thus increasing the amount of data obtained in response to each treatment, as compared to a simple between-group design), and 2) in a properly designed crossover study, there is no (or a highly reduced) systematic relation (confounding) between the treatments and the order in which they are administered.

A disadvantage is that the potential carry-over effects (different responses of experimental units to the same treatment depending on the order of the preceding treatments) cannot be fully identified or completely controlled as long as all possible orders of the treatments are not administered and the sample size of each order of treatments is not large enough (which may be impractical, since it might require very large numbers of observations).