Comparison of approaches to constructing confidence intervals for mediating effects using structural equation models
Comparison of approaches to constructing confidence intervals for mediating effects using structural equation models
复制标题
DOI：
10.1080/10705510709336745
复制
发表时间：
2007-01-01
影响因子：
6
通讯作者：
Cheung, Mike W. -L.
中科院分区：
文献类型：
--
作者：
Cheung, Mike W. -L.
中介变量是解释自变量和因变量之间关联的变量。结构方程模型（SEM）广泛用于测试具有中介效应的模型。本文说明了如何为 SEM 中的各种模型构建中介效应的置信区间 (CI)。具体来说，说明了具有 I 中介者、2 个中间中介者、2 个特定中介者和 2 个独立组中的 I 中介者的中介模型。通过使用虚拟变量（Rindskopf，1984），可以使用一些现有的 SEM 软件包（例如 LISREL、Mplus 和 Mx）轻松构建 Wald CI、百分位数引导 CI、偏差校正引导 CI 以及基于似然性的中介效应 CI。蒙特卡罗模拟研究用于比较这些 CI 的覆盖概率。结果表明，当中介效应较大或样本量较大时，这些 CI 的覆盖概率具有可比性。然而，当中介效应和样本量都很小时，Bootstrap CI 和基于似然的 CI 优于 Wald CI。讨论了这种 SEM 方法对未来研究的扩展。
Mediators are variables that explain the association between an independent variable and a dependent variable. Structural equation modeling (SEM) is widely used to test models with mediating effects. This article illustrates how to construct confidence intervals (CIs) of the mediating effects for a variety of models in SEM. Specifically, mediating models with I mediator, 2 intermediate mediators, 2 specific mediators, and I mediator in 2 independent groups are illustrated. By using phantom variables (Rindskopf, 1984), a Wald CI, percentile bootstrap CI, bias-corrected bootstrap CI, and a likelihood-based CI on the mediating effect are easily constructed with some existing SEM packages, such as LISREL, Mplus, and Mx. Monte Carlo simulation studies are used to compare the coverage probabilities of these CIs. The results show that the coverage probabilities of these CIs are comparable when the mediating effect is large or when the sample size is large. However, when the mediating effect and the sample size are both small, the bootstrap CI and likelihood-based CI are preferred over the Wald CI. Extensions of this SEM approach for future research are discussed.