Integrating Moderation and Mediation
Shruti R. Sardeshmukh, Robert J. Vandenberg
Abstract
Shruti R. Sardeshmukh, Robert J. Vandenberg
Abstract
It is increasingly common to test hypotheses combining moderation and mediation. Structural equation modeling (SEM) has been the favored approach to testing mediation hypotheses. However, the biggest challenge to testing moderation hypotheses in SEM was the complexity underlying the modeling of latent variable interactions. We discuss the latent moderated structural equation procedure (LMS) approach to specifying latent variable interactions, which is implemented in Mplus, and offer a simple and accessible way of testing combined moderation and mediation hypotheses using SEM. To do so, we provide sample code for six commonly encountered moderation and mediation cases and relevant equations that can be easily adapted to researchers’ data. By articulating the similarities in the two different approaches, discussing the combination of moderation and mediation, we also contribute to the research methods literature.
OpenAlex reports 301 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
It is increasingly common to test hypotheses combining moderation and mediation. Structural equation modeling (SEM) has been the favored approach to testing mediation hypotheses. However, the biggest challenge to testing moderation hypotheses in SEM was the complexity underlying the modeling of latent variable interactions. We discuss the latent moderated structural equation procedure (LMS) approach to specifying latent variable interactions, which is implemented in Mplus, and offer a simple and accessible way of testing combined moderation and mediation hypotheses using SEM. To do so, we provide sample code for six commonly encountered moderation and mediation cases and relevant equations that can be easily adapted to researchers’ data. By articulating the similarities in the two different approaches, discussing the combination of moderation and mediation, we also contribute to the research methods literature.
Key concepts: Moderation, Structural equation modeling, Mediation, Latent variable, Psychology, Moderated mediation, Latent variable model, Test (biology)