Understanding Structural Equation Modeling: An Introduction for Social Work Researchers

Significance

Structural equation modeling (SEM) is a flexible and powerful statistical analysis framework that is appropriate for much of the data used by social work researchers. SEM permits the sophisticated evaluation of social work scales and instruments, as well as the testing of hypotheses about complex developmental processes and intervention effects. Because it requires specialized software and a knowledge base that goes beyond the statistical training of many social workers researchers, SEM is often either avoided or misused. This workshop is geared toward the many social work researchers who have the desire to expand their statistical expertise and improve the quality of their publications by becoming familiar with the basics of SEM. Recent SSWR conferences have featured workshops on advanced SEM techniques (e.g., multiple group testing), but have not offered an option for those who seek a general introduction to SEM.

Content

This SEM workshop is designed for social work faculty, researchers, and students who view themselves primarily as practitioners or substantive experts, but who nonetheless desire to use the statistical tools that are best suited for the complex data and research questions of their topic areas. Workshop leaders will assume that participants have general knowledge of regression, correlation, and comparison of group means. The workshop will comprise five components: (a) a brief review of core statistical concepts, (b) an overview of the SEM framework and how it differs from more familiar approaches, (c) an interactive discussion of the role of theory in two major types of SEM models—confirmatory factor models and general structural models, (d) a presentation and handouts on the steps of conducting and evaluating SEM models, and (e) information on what journal reviewers expect to see in an SEM article. The main focus will be on understanding the conceptual underpinnings of SEM and not on conducting SEM analyses. However, a simple model will be demonstrated and participants will learn how to interpret SEM output. The example will make use of Mplus software. The workshop will conclude with a discussion of next steps for participants in developing their SEM expertise.

Conclusion

This workshop fills an important need in SSWR’s Research Design and Measurement (RD&M) cluster. Attendees of this workshop will gain knowledge and practice in three key areas: (a) basic knowledge about SEM and its potential as a research tool, (b) understanding of the steps for conducting an SEM analysis and interpreting its result, and (c) understanding of key elements of SEM analyses that must be reported in high-quality journals. Participants will also leave with a list of resources and suggestions for improving their SEM skills going forward. Materials provided to participants will include information on the steps in conducting an SEM and interpreting output, annotated examples of SEM articles’ methods sections, a list of SEM resources, and annotated SEM output from Mplus 7.3.