Crossover experiments

Crossover designs and within‐subject designs are designs where each experimental unit receives a series of treatments over time. Because treatments are assigned at different time points, one must worry about whether there is carryover or residual effects from a treatment given in a previous period to a subsequent period. This article introduces crossover designs and discusses the statistical analysis of crossover designs. When one has more than two periods and/or more than three treatments, then special designs can be obtained where one can measure a direct comparison between the treatments even if carryover effects exist.