贝叶斯因子序列分析:实验设计中平衡信息与效率的新方法  

Sequential Bayes Factor Analysis:Balance Informativeness and Efficiency in Designing Experiments

作　　者：郑元瑞 胡传鹏 ZHENG Yuan-rui;HU Chuan-peng(School of Psychology,Nanjing Normal University,Nanjing 210097,China;Faculty of Education,Kunming City College,Kunming 650106,China)

机构地区：[1]南京师范大学心理学院,南京210097 [2]昆明城市学院教育学院,昆明650106

出　　处：《应用心理学》2024年第2期158-171,共14页Chinese Journal of Applied Psychology

摘　　要：实验设计的关键是平衡信息量与效率。贝叶斯因子序列分析利用贝叶斯因子不断更新证据的特点,通过贝叶斯因子标准和在收集数据过程中的序列分析来平衡信息量与效率。本文展示如何使用开源软件JASP和R实现该分析的三个步骤:确定关键效应、确定停止标准、数据收集中序列分析并决策。该方法考虑现实条件且简单易行,可帮助研究者更有效地进行实验。The key of experimental design is to balance between informativeness and efficiency.However,power analysis only focuses on informativeness and is difficult to implement.Here,sequential Bayes factor analysis takes the advantage of Bayes factor’s ability to continuously update the evidence and reach a trade-off between informativeness and efficiency by setting Bayes factor criteria and the sequential analysis during data collection.The present primer demonstrates how to perform three steps of sequential Bayes factor analysis using the open-source software JASP and R.This method considers practical issues in real research practices and is easy to implement,which can help researchers to design more efficient experiments.

关 键 词：贝叶斯因子序列分析 统计检验力 实验设计 JASP R 

分 类 号：B849[哲学宗教—应用心理学]