NEW TABLES FOR MULTIPLE COMPARISONS WITH CONTROL
NEW TABLES FOR MULTIPLE COMPARISONS WITH CONTROL
复制标题
DOI：
10.2307/2528490
复制
发表时间：
1964-01-01
期刊：
影响因子：
1.9
通讯作者：
DUNNETT, CW
中科院分区：
文献类型：
--
作者：
DUNNETT, CW
表中给出了几种处理平均值与对照或标准处理平均值之间的多重比较检验（使用t统计量），其中需要对处理和对照之间的差异进行显著性检验或置信限检验。所提出的程序说明了一个例子有关的药物治疗对公鸡胸肌脂肪含量的影响。表格给出了1 - 20个处理比较达到联合置信系数0.95或0.99的临界值，并包括对照组相对于处理组的观察值数量的调整因子（如果这些不相等）。
Tables are given for multiple comparisons tests, using t-statistics, between several treatment means and a control or standard treatment mean, in which either tests of significance or confidence limits on the differences between treatments and control are required. The presented procedure is illustrated by an example concerned with effects of drug treatments on fat content of breast muscle in cockerels. Tables give critical values to achieve a joint confidence coefficient of .95 or .99 for 1 to 20 treatment comparisons, and include adjustment factors for the number of observations on control relative to a treatment if these are unequal.