贝叶斯估计在小样本潜变量中介效应研究中的优势分析*

高菲菲 ,  马洪林 ,  孟佳 ,  田延森 ,  宋奕德 ,  刘洋

中国卫生统计 ›› 2026, Vol. 43 ›› Issue (3) : 364 -368.

PDF (466KB)
中国卫生统计 ›› 2026, Vol. 43 ›› Issue (3) : 364 -368. DOI: 10.11783/j.issn.1002-3674.2026.03.008
论著

贝叶斯估计在小样本潜变量中介效应研究中的优势分析*

作者信息 +

Advantages of Bayesian Estimation in the Study of Mediating Effects of Small Sample Latent Variables

Author information +
文章历史 +
PDF (476K)

摘要

目的 本研究以一项基于中学生家庭功能和手机依赖的横断面现场调查数据为实证案例,旨在验证贝叶斯结构方程模型在小样本潜变量中介效应分析中的优势。方法 采用极大似然法和有/无信息先验的贝叶斯估计分别对不同样本量的实证案例数据进行建模,并比较结果的稳定性和准确性。结果 在分析小样本时,贝叶斯估计易收敛,拟合效果良好,尤其是有信息先验的贝叶斯估计中,估计出来的中介效应指标更精确;在分析大样本时,研究假设路径均成立,有信息先验模型的贝叶斯估计优于无信息先验模型。结论 在条件允许下,尽可能获得大样本数据进行分析,以保证研究结果的检验力和准确性。而在无法获得大样本数据时,有信息先验的贝叶斯估计应是首选方案。

Abstract

Objective Using a cross-sectional field survey data based on family function and mobile phone dependency of middle school students as an empirical case, this study aims to verify the advantages of Bayesian structural equation model in the analysis of the mediating effects of small sample latent variables.Methods The maximum likelihood method and Bayesian estimation with/without information prior were used to model the empirical case data of different sample sizes respectively, and the stability and accuracy of the results were compared.Results When analyzing small samples, Bayesian estimation is easy to converge and has good fitting effect. Especially in Bayesian estimation with information priori, the estimated intermediary effect indicators are more accurate. When analyzing large samples, the research hypothesis paths are all valid. The Bayesian estimation with information prior model is superior to the model without information prior. Conclusion If conditions permit, large sample data can be obtained for analysis as far as possible to ensure the testing power and accuracy of the research results. When large sample data cannot be obtained, Bayesian estimation for small sample analysis should be the first choice.

关键词

结构方程模型 / 贝叶斯估计 / 先验分布 / 最大似然法

Key words

Structural equation model / Bayesian estimation / Prior distribution / Maximum likelihood method

引用本文

引用格式 ▾
高菲菲,马洪林,孟佳,田延森,宋奕德,刘洋. 贝叶斯估计在小样本潜变量中介效应研究中的优势分析*[J]. 中国卫生统计, 2026, 43(3): 364-368 DOI:10.11783/j.issn.1002-3674.2026.03.008

登录浏览全文

4963

注册一个新账户 忘记密码

参考文献

[1]

van de Schoot R, Winter SD, Ryan O, et al. A systematic review of Bayesian articles in psychology: The last 25 years[J].Psychol Methods, 2017,22(2):217-239.

[2]

Button KS, Ioannidis JP, Mokrysz C, et al. Power failure: why small sample size undermines the reliability of neuroscience[J].Nature Reviews Neuroscience, 2013,14(5):365-376.

[3]

Baldwin SA, Fellingham GW. Bayesian methods for the analysis of small sample multilevel data with a complex variance structure[J].Psychol Methods, 2013,18(2):151-164.

[4]

Depaoli S, van de Schoot R. Improving transparency and replication in Bayesian statistics: The WAMBS—Checklist[J].Psychol Methods,2017,22(2):240-261.

[5]

McNeish D, Stapleton LM. Modeling Clustered Data with Very Few Clusters[J].Multivariate Behav Res, 2016,51(4):495-518.

[6]

Boomsma A. The robustness of maximum likelihood estimation in structura equation models[J].Psychometrika, 2014,16(55): 80-88.

[7]

van de Schoot R, Broere JJ, Perryck KH, et al. Analyzing small data sets using Bayesian estimation: the case of posttraumatic stress symptoms following mechanical ventilation in burn survivors[J].Eur J Psychotraumatol, 2015,6:25216.

[8]

MacCallum RC, Edwards MC, Cai L. Hopes and cautions in implementing Bayesian structural equation modeling[J].Psychol Methods, 2012,17(3):340-353.

[9]

Young ME. Bayesian data analysis as a tool for behavior analysts[J].Exp Anal Behav, 2019,111(2): 225-238.

[10]

张沥今,陆嘉琦,魏夏琰,等. 贝叶斯结构方程模型及其研究现状[C]//中国心理学会. 第二十二届全国心理学学术会议摘要集. 中山大学心理学系,浙江大学心理与行为科学系,2019:1.

[11]

朱慧明, 韩玉启. 贝叶斯多元统计推断理论[M].北京:科学出版社, 2006.

[12]

Smilkstein G, Ashworth C, Montano D. Validity and reliability of the family APGAR as a test of familyfunction[J].The Journal of Family Practice, 1982,15 (2) :303-311.

[13]

吕繁,顾湲. 家庭APGAR 问卷及其临床应用[J].国外医学(医院管理分册),1995(2):56-59.

[14]

Leung L. Linking Psychological Attributes to addiction and Improper Use of the Mobile Phone among adolescents in hong kong[J].Children & Media, 2008,2(2): 93-113.

[15]

Rosenberg FR, Rosenberg M, McCord J. Self—esteem and delinquency[J].Youth Adolesc, 1978,7(3):279-294.

[16]

Gelman, Carlin JB, Stern, et al. Bayesian data analysis[J].Wiley Interdisciplinary Reviews Cognitive Science, 2014,1(5): 658-676.

[17]

Smid SC, Winter SD. Dangers of the Defaults: A Tutorial on the Impact of Default Priors When Using Bayesian SEM With Small Samples[J].Front Psychol, 2020,11:611963.

[18]

Ulitzsch E, Lüdtke O, Robitzsch A. Alleviating estimation problems in small sample structural equation modeling—A comparison of constrained maximum likelihood, Bayesian estimation, and fixed reliability approaches[J].Psychol Methods, 2023,28(3):527-557.

[19]

Muthén B, Asparouhov T. Bayesian structural equation modeling: a more flexible representation of substantive theory[J].Psychol Methods, 2012,17(3):313-335.

[20]

van de Schoot R, Kaplan D, Denissen J, et al. A gentle introduction to Bayesian analysis: applications to developmental research[J].Child Dev, 2014,85(3):842-860.

基金资助

辽宁省社会科学规划基金(L20CRK001)

AI Summary AI Mindmap
PDF (466KB)

0

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉