with Gregory R. Hancock
Structural equation modeling, in general, represents a process for assessing theories about causal relations among measured and latent variables through modeling those theories’ implications for the variances and covariances of observed variables. Mean structure models extend these models to allow for the assessment of theories about population differences in latent means, akin to analysis of variance on latent outcomes. The ability to do so greatly enhances researchers’ capacity to deal with experimental and quasi-experimental settings, facilitating more statistical power and/or requiring fewer subjects to conduct such studies. Latent growth modeling is a special application of structural equation modeling, which may or may not involve a mean structure, that allows one to assess individual differences in subjects’ growth/change over time in one or more measured or latent outcomes, as well as the measured and latent time- independent and time-dependent determinants of that change. This invited lecture will briefly review structural equation modeling, and then provide an introduction to latent mean and latent growth models within a structural equation modeling framework.