This workshop will illustrate several important advances in structural equation modeling (SEM). Emphasis will be on a new graph-theoretic implementation of SEM, which permits both likelihood and Bayesian estimation methods, as well as non-Gaussian responses and nonlinear linkage functions. The workshop will provide (1) empirical demonstrations to highlight how these advances increase SEM’s capacity for retrospective reasoning and prediction; and (2) guided exploration of tutorials that are designed to increase participant’s confidence in applying the new tools. Participants must have a basic understanding of SEM, background in modern statistics, and general competence with the R statistical platform.