This workshop will illustrate several important advances in structural equation modeling (SEM) and is intended for those with at least a basic background in the methodology. Emphasis will be on a graph-theoretic implementation of SEM that permits non-Gaussian responses, nonlinear linkage functions, and multi-level data. Also included will be a discussion of formal causal analysis methods. The workshop will present both principles and empirical demonstrations. I will also highlight how the advances increase SEM's capacity for retrospective and prospective reasoning. Participants must have a basic understanding of SEM, background in modern statistics, and general competence with the R statistical platform.