INCOMPLETE BLOCK DESIGNS
INCOMPLETE BLOCK DESIGNS
复制标题
DOI：
10.1201/9781315136318-11
复制
发表时间：
2003
期刊：
影响因子：
--
通讯作者：
D. Weber;J. Skillings
中科院分区：
文献类型：
--
作者：
D. Weber;J. Skillings
采用完整的区组设计，所有的处理都代表了实验的每个区块。有时你可能有大量的治疗方法要比较，而你可用的模块只能处理有限数量的这些治疗方法。在这些情况下，使用不完整的区块设计。这些设计是部分混淆的设计，允许一些信息和所有对比，只要有足够的复制。一般说来，每一类都可以用不同的信息量进行估计，尽管有些设计具有某种类型的平衡。具有某种平衡的不完全区组设计的一种类型是平衡不完全区组设计(BIB)。一般来说，不完全区组设计可以被认为是一个双因素试验(区组和处理)，具有不同的数量，并且区组和处理因素之间没有交互作用，导致以下线性模型：
With complete block designs, all treatments are represented in each block of the experiment. Sometimes you may have a large number of treatments that you want to compare and the blocks that you have available can only handle a limited number of these treatments. Incomplete block designs are used in these situations. These designs are partially confounded designs that allow some information an all contrasts, as long as there is adequate replication. In general, each class may be estimated with a different amount of information, although there are designs that have some type of balance. One type of incomplete block design that has some balance is the balanced incomplete block design (BIB). In general, an incomplete block design can be thought of as a two-factor experiment (block and treatment) with unequal numbers and no interaction between the block and treatment factor, leading to the following linear model: