Statistical reverberation theory is a foundational framework for describing sound behavior in enclosures but often falls short in practical settings, particularly in rooms with non-compact geometries or inhomogeneous material distributions. This study examines the applicability of statistical reverberation theory, focusing on absorption and scattering coefficients that govern the build-up and decay of non-specular sound fields. Motivated by challenges in the hybrid modeling framework of Wittebol et al. (2024), this research evaluates the theory’s reliability under varying geometric and material conditions. Absorption and scattering coefficients are acquired through simulations using the image source method for cuboid enclosures, followed by Schroeder integration and exponential fitting. These are compared to the area-weighted coefficients assumed by statistical theory. Results demonstrate that, regarding room shape, compact geometries yield better agreement with theoretical assumptions, while regarding material distribution, scenarios with opposing walls, each with uniform properties but differing between sets, show the poorest agreement. These findings highlight the importance of evaluating room compactness and material distribution before relying on statistical reverberation theory. By assessing the limitations of statistical reverberation theory in non-homogeneous or non-compact settings, this study offers insights into when hybrid models can effectively incorporate elements of statistical theory to describe non-specular sound fields.