Master Methodologists
Integrating
Mediation and Moderation Analysis
Andrew Hayes, PhD, The Ohio State University, Columbus, Ohio, USA
Saturday, November 3
9:00 a.m. – 11:45 a.m.
New
Developments in Latent Variable Modeling: Multilevel and Mixture
Analysis
Bengt O. Muthén, PhD, University of California, Los Angeles, California, USA
Saturday, November 3
3:00 p.m. – 4:15 p.m.
Saturday, November 3, 9:00 a.m. – 11:45 a.m.
Integrating Mediation and Moderation Analysis
Andrew Hayes, PhD, The Ohio State University, Columbus, Ohio,
USA
Primary Keyword: Research Methodology
Presentation Level: Intermediate
Region: Global
As research in any particular area develops and evolves, attention naturally shifts away from establishing the existence of some kind of causal effect between two variables to understanding how the effect operates (mediation) and when the effect exists or is strong versus when it is absent or weak (moderation). Few would dispute that all effects exist through some kind of mechanism, and all effects have boundary conditions. Thus, an analysis which attempts to answer only how or when but not both is necessarily incomplete in significant ways. Recently, methodologists have been describing approaches to integrating moderation and mediation analysis into a single integrated model.
This session has four basic objectives: to introduce the audience to the theoretical and substantive rationale for combining moderation and mediation analysis in a single integrated statistical model, (2) to provide a tutorial on some of the basic statistical concepts including modern approaches to inference, (3) to illustrate by example how such analytical integration has been undertaken in some existing published research, and (4) to demonstrate an easy-to-use statistical tool developed for SPSS and SAS that makes this analytical approach extremely simple to conduct.
Andrew F. Hayes (PhD, Cornell University; BA, San Jose State University) is associate professor of psychology and associate professor of communication at The Ohio State University, where he teaches data analysis and research design. His research on statistical methods has been published in such places as Psychological Methods, Multivariate Behavioral Research, Behavior Research Methods, and the Journal of Educational and Behavioral Statistics. He conducts workshops on moderation and mediation analysis throughout North America, Europe, and the Middle East, and he has authored a book on this topic being published by Guilford Press in 2013.
Saturday, November 3, 3:00 p.m. – 4:15 p.m.
New Developments in Latent Variable Modeling: Multilevel and
Mixture Analysis
Bengt O. Muthén, PhD, University of
California, Los Angeles, California, USA
The use of latent variables is a common theme in many statistical analyses. Continuous latent variables appear not only as factors measured with errors in factor analysis, item response theory, and structural equation modeling, but also appear in the form of random effects in growth modeling, components of variation in complex survey data analysis and multilevel modeling, frailties and liabilities in survival and genetic analyses, latent response variables with missing data, priors in Bayesian analysis, and as counterfactuals and potential outcomes in causal analysis. In addition, categorical latent variables appear as latent classes in finite mixture analysis and latent transition analysis (Hidden Markov modeling), latent trajectory classes in growth mixture modeling, and latent response variables with missing data on categorical variables.
Understanding the unifying theme of latent variable modeling provides a way to break down barriers between seemingly disparate types of analyses. Researchers need to be able to move freely between analysis types to more easily answer their research questions. To provide answers to the often complex substantive questions, it is also fruitful to use latent variable techniques to combine different analysis types. This talk discusses examples that use combinations of multilevel, latent class, and longitudinal modeling features in the new Mplus Version 7.