A simple test for variance effects in experimental designs.
A simple test for variance effects in experimental designs.
复制标题
实验设计中方差效应的简单检验。
DOI：
10.1037/0033-2909.89.3.570
复制
发表时间：
1981
影响因子：
22.4
通讯作者：
R. O’brien
中科院分区：
文献类型：
--
作者：
R. O’brien
尽管实验效果通常是通过组均值对比来评估的，但在某些情况下，组方差之间的差异也令人感兴趣。此类分析很少用于行为研究，可能是因为最常见的方法对于非正态分布的数据并不稳健。提出了一种程序，通过使用因变量的变换执行常规方差分析来生成单元格方差相等性的稳健检验。讨论了特殊对比（例如，简单效果、子效果）并给出了示例。（29 ref）（PsycINFO 数据库记录 (c) 2016 APA，保留所有权利）
Although experimental effects are usually assessed through contrasts of group means, there are situations in which differences among the groups' variances are also of interest. Such analyses are infrequently used in behavioral research, possibly because the most common methods are not robust to nonnormally distributed data. A procedure is presented that produces robust tests of the equality of cell variances by performing a regular ANOVA using a transformation of the dependent variable. Special contrasts (eg, simple effects, subeffects) are discussed and an example is given.(29 ref)(PsycINFO Database Record (c) 2016 APA, all rights reserved)