您当前的位置: > 详细浏览

贝叶斯方差分析在JASP中的实现

请选择邀稿期刊:

The implementation of Bayesian ANOVA in JASP: A practical primer

摘要: 贝叶斯统计应用于假设检验的方法——贝叶斯因子——在心理学研究中的应用日渐增加。贝叶斯因子能分别量化所支持的相应假设或模型的证据,进而根据其数值大小做出当前数据更支持哪种假设或模型的判断。然而,国内尚缺乏对方差分析的贝叶斯因子的原理与应用的介绍。基于此,本文首先介绍贝叶斯方差分析的基本思路及计算原理,并结合实例数据,展示如何在JASP中对五种常用的心理学实验设计(单因素组间设计、单因素组内设计、二因素组间设计、二因素组内设计和二因素混合设计)进行贝叶斯方差分析及如何汇报和解读结果。贝叶斯方差分析提供了一个能有效替代传统方差分析的方案,是研究者进行统计推断的有力工具。
Abstract: The application of Bayesian statistics to hypothesis testing - Bayes factors - is increasing in psychological science. Bayes factors quantify the evidence supporting the competing hypothesis or model, respectively, thereby making a judgment about which hypothesis or model is more supported by the data based on its value. The principles and applications of Bayes factor for ANOVA are, however, not available in China. We first present the theoretical foundation of Bayesian ANOVA and its calculation rules. It also shows how to perform Bayesian ANOVA and how to interpret and report the results of five common designs (one-factor between-group design, one-factor within-group design, two-factor between-group design, two-factor within-group design, and two-factor mixed design) using example data. Theoretically, Bayesian ANOVA is an effective alternative to conventional ANOVA as a powerful vehicle for statistical inferences.

版本历史

[V3] 2023-09-18 10:44:50 ChinaXiv:202209.00140V3 下载全文
[V1] 2022-09-27 21:38:07 ChinaXiv:202209.00140v1 查看此版本 下载全文
点击下载全文
预览
同行评议状态
通过
许可声明
metrics指标
  •  点击量13000
  •  下载量929
评论
分享