有序处理连续比较的置信区间

为有序均值的连续比较提出了一个新的检验过程.他们的过程对满足简单序的均值有重要的应用价值,例如在研究增长剂量对药物效用的影响. 与其它检验过程相比(例:Hayterucite{4}中的检验), 其优点在于产生了更短的连续比较的置信限,从而能够提供更多机会发现在何药剂量处有不同的效用.但作为有序均值的齐性检验, 它的势表现远劣于其它检验.本文的目的是提出一检验过程在尽量保持Lee andSpurrierucite{8}检验的优点的同时大大地提高其势表现.

Confidence Intervals for Successive Comparisons of Ordered Treatment Effects

present a new procedure formaking successive comparisons between ordered treatments. Theirprocedure has important applications for problems where thetreatments can be assumed to satisfy a simple ordering, such as fora sequence of increasing dose-levels of a drug. The advantage oftheir procedure is that it provides more chance to detect whenchanges in the treatment means occur than other test procedures (forexample: test in Hayterucite{4}). The disadvantage of theirprocedure is that it is not as powerful as other test procedures. Inthis paper we propose a test procedure which try to keep theadvantage of Lee and Spurrier's procedure and promote the powerperformance of their test procedure.