Optimal Crossover Designs for the Comparison of Two Treatments in the Presence of Carryover Effects and Autocorrelated Errors
残留效应与自相关误差下比较两种处理的最优交叉设计
A model for responses in a crossover experiment that includes direct and first-order carryover treatment effects, together with an autocorrelated error term, is considered when there are two treatments. A method for generating designs that minimize the variance of the estimated treatment effects is proposed and explicit results are given for the three and four period cases. Efficiencies of some practical designs are calculated.