Switchback Trials for More than Two Treatments

Summary To take better advantage of the high sensitivity of switch-back or double-reversal trials commonly used to compare two treatments and which have been extended to permit the comparison of three or more treatments, certain convenient and useful features have been added. Designs are given for 3, 4, 5, 6, 7, and 9 treatments. The statistical analysis is outlined symbolically and is illustrated numerically with uniformity data. Missing value formulas are given.