Institute for Health and Sport, Victoria University, Melbourne, Australia.
School of Health and Biomedical Sciences, RMIT University, Carlton, Australia.
BMC Psychiatry. 2023 Jul 28;23(1):546. doi: 10.1186/s12888-023-05028-9.
Confirmatory Factor Analysis (CFA) has been a popular yet limited approach to assessing latent factor structures. Despite items rarely loading exclusively on one latent factor in multifactorial scales, CFA assumes all indicators/items should load uniquely on their allocated latent dimensions. To address this weakness, Exploratory Structural Equation Modeling (ESEM) combines exploratory factor analyses (EFA) and CFA procedures, allowing cross-loadings to occur when assessing hypothesized models. Although such advantages have enhanced ESEM popularity, its adoption is often limited by software rigidity and complex coding difficulties. To address these obstacles, the current tutorial presents a streamlined, step-by-step approach using the open-source software R while providing both R and Mplus ESEM syntax. The tutorial demonstrates the sequence of the ESEM stages by examining the frequently debated Strengths and Difficulties Questionnaire (SDQ) factor structure, using openly accessible data from the Longitudinal Study of Australian Children (LSAC). As ESEM may allow a better understanding of the complex associations in multidimensional scales, this tutorial may optimize the epidemiological and clinical assessment of common yet multifaceted psychiatric presentations.
验证性因子分析(CFA)是一种流行但有限的方法,用于评估潜在因子结构。尽管多因子量表中的项目很少只加载到一个潜在因子上,但 CFA 假设所有指标/项目都应唯一加载到其分配的潜在维度上。为了解决这个弱点,探索性结构方程建模(ESEM)结合了探索性因子分析(EFA)和 CFA 程序,允许在评估假设模型时发生交叉加载。尽管这些优势提高了 ESEM 的流行度,但由于软件的刚性和复杂的编码困难,其采用往往受到限制。为了解决这些障碍,本教程使用开源软件 R 提供了一种简化的、逐步的方法,同时提供了 R 和 Mplus 的 ESEM 语法。该教程通过使用来自澳大利亚儿童纵向研究(LSAC)的公开可用数据,检查经常讨论的困难问卷(SDQ)因子结构,演示了 ESEM 阶段的顺序。由于 ESEM 可以更好地理解多维量表中的复杂关联,因此本教程可以优化常见但多方面的精神科表现的流行病学和临床评估。