华东师范大学学报(教育科学版)

华东师范大学学报(教育科学版) ›› 2023, Vol. 41 ›› Issue (1): 1-15.doi: 10.16382/j.cnki.1000-5560.2023.01.001

• 特稿 •    

潜在类别模型的原理、步骤及程序

温忠麟1, 谢晋艳1, 王惠惠2   

  1. 1. 华南师范大学心理学院/心理应用研究中心，广州 510631
    2. 宁夏大学教育学院，银川 750021
  • 出版日期:2023-01-01 发布日期:2022-12-29
  • 基金资助:
    国家自然科学基金项目“纵向因果分析中使用工具变量应对内生性问题的建模策略及其应用”(32171091)

Principles, Procedures and Programs of Latent Class Models

Zhonglin Wen1, Jinyan Xie1, Huihui Wang2   

  1. 1. School of Psychology/Center for Studies of Psychological Application, South China Normal University, Guangzhou 510631, China
    2. School of Education, Ningxia University, Yinchuan 750021, China
  • Online:2023-01-01 Published:2022-12-29

RichHTML

301

PDF

3905

摘要：

潜在类别分析和潜在剖面分析相应的模型统称为潜在类别模型，它是根据个体在观测指标上的不同反应模式将其进行分类，从而达到识别群体异质性的一类统计方法，在教育学、心理学等社科领域日益受到应用研究者的关注。然而现有的中文文献对此类模型的统计原理和分析步骤的介绍不易为教育研究工作者所接受。本文系统讲述潜在类别模型的基础知识、统计原理、分析步骤和Mplus程序，厘清潜在类别模型的后续分析所涉及的多种方法和选用策略，有助于增进应用研究者对潜在类别模型的原理理解和方法掌握，推动潜在类别模型在教育研究中的应用。

关键词: 潜在类别模型, 潜在类别分析, 潜在剖面分析, 潜在转变分析, 类别概率, 条件概率

Abstract:

The models used in Latent Class Analysis and Latent Profile Analysis are collectively referred to as latent class models, a kind of statistical methods of classifying individuals according to their different response patterns in observation indicators, so as to identify population heterogeneity. It has attracted increasing attention from applied researchers in the fields of pedagogy, psychology, and other social science disciplines. However, it is not easy for most education researchers to understand the existing Chinese literature on the statistical principles and analytical procedures of such models. This paper systematically introduces the basic knowledge, statistical principles, analytical procedures and Mplus programs of latent class models, and clarifies various methods and selection strategies involved in the subsequent analysis of these models. It would help applied researchers enhance their understanding of the principles and methods of the latent class models, and promote the application of these models to educational research.

Key words: latent class model, Latent Class Analysis, Latent Profile Analysis, Latent Transition Analysis, category probability, conditional probability