贝叶斯推论统计在心理学研究中的应用

Abstract

近年来贝叶斯统计学越来越受到心理学界的关注。该方法的基本逻辑是综合先验信息和实验结果得出一个后验概率，令研究者可以直接地、客观地检验研究假设，或使用贝叶斯因子比较哪种研究假设能更好解释实验数据。本文总结了这种方法的优势和劣势，举例说明如何在心理学研究中应用它，并介绍了可以使用的软件和教材，以此作为对当前心理学研究中统计方法的一种补充。 In recent years, the Bayesian Data Analysis method has increasingly received attention from psychological researchers. This method allows them to directly and objectively estimate the probability of a research hypothesis by deriving a posterior probability from the prior distribution and their own research data, or to use the Bayesian factor to directly compare which of two hypotheses can better explain their data. In this literature review paper, we summarized several major strengths of this inference method and its deficiencies. We used examples to illustrate how it can be used in psychological research and summarized some software and textbooks that psychological researchers can use to learn the Bayesian data analysis method. This article aims to introduce a supplementary method into psychological research.