Kai Cortina, Professor of Psychology, will present on best practices for reporting results of Structural Equation Models. Structural Equation Modeling has evolved into a standard statistical procedure with applications way beyond its initial focus (path analysis with latent variables). A long and somewhat confusing list of model fit statistics were developed over the last 3 decades. In his talk, Kai will briefly present the rationale behind most of them and provide some guidance for a proper use with focus on best practices for submitting manuscripts to APA journals. The group will discuss: (a) What is better: strong statistics with strong assumptions (e.g., chi_square) or soft statistics with few assumptions (e.g., AIC, BIC)? (b) How useful are fit comparisons across models that are not hierarchically nested?