An assessment of statistical methods for non‐independent data in ecological meta‐analyses: Reply
An assessment of statistical methods for non‐independent data in ecological meta‐analyses: Reply
复制标题
生态荟萃分析中非独立数据统计方法的评估：回复
DOI：
10.1002/ecy.3578
复制
发表时间：
2021
期刊：
影响因子：
4.8
通讯作者：
Bence, James R.
中科院分区：
文献类型：
--
作者：
Song, Chao;Peacor, Scott D.;Osenberg, Craig W.;Bence, James R.
最近，Nakagawa等人(2021)对我们关于生态荟萃分析中非独立数据的统计方法的论文(Song等人)提出了及时和有见地的评论。2020)。他们的评论强调了在Meta分析中使用分层模型来解决非独立性的价值，并提出了两个断言：(1)首先计算每篇论文的加权平均效应大小，然后在随机效应模型中分析论文平均值的两步法应用范围有限；(2)分层模型中避免I型错误率膨胀的几种解决方案已经存在，并且可以用R。
Recently, Nakagawa et al.(2021) provided a timely and insightful comment to our paper on statistical methods for non-independent data in ecological metaanalyses (Song et al. 2020). Their comment highlighted the value of using hierarchical models in meta-analysis to address non-independence, and offered two assertions:(1) that a two-step method that first calculates a weighted mean effect size of each paper and then analyzes the paper mean in a random effect model has limited scope of application and (2) that several solutions to avoid inflated type I error rates in hierarchical models already exist and can be implemented with existing software packages in R.