Multiplicity Control, School Uniforms, and Other Perplexing Debates
Multiplicity Control, School Uniforms, and Other Perplexing Debates
复制标题
DOI：
10.1037/cbs0000075
复制
发表时间：
2017-07-01
影响因子：
2.5
通讯作者：
Cribbie, Robert A.
中科院分区：
文献类型：
--
作者：
Cribbie, Robert A.
心理学研究人员经常面临同时分析多个关系的问题。例如，这可能涉及回归框架中的多个结果变量或多个预测变量。当前的建议通常会引导研究人员对家庭或错误的I型误差控制率限制错误拒绝零假设的可能性。建议逐步修改的 - 邦弗罗尼程序遵循此建议。然而，长期以来反对多样性控制的论点与现代假设意义检验的现代厌恶相结合，有必要重新审视这一辩论。本文是对多重控制辩论的两面的探索，其目的是教育有关当事方有关进行多个相关测试的最佳实践。
Researchers in psychology are frequently confronted with the issue of analysing multiple relationships simultaneously. For example, this could involve multiple outcome variables or multiple predictors in a regression framework. Current recommendations typically steer researchers toward familywise or falsediscovery rate Type I error control to limit the probability of incorrectly rejecting the null hypothesis. Stepwise modified-Bonferroni procedures are suggested for following this recommendation. However, longstanding arguments against multiplicity control combined with a modern distaste for null hypothesis significance testing have warranted revisiting this debate. This paper is an exploration of both sides of the multiplicity control debate, with the goal of educating concerned parties regarding best practices for conducting multiple related tests.