Computational modeling of systems governed by non-bonded interactions, especially van der Waals (dispersion) interactions, is currently a difficult task, since many conventional quantum mechanical techniques neglect such interactions. For example, the popular semi-empirical and Hartree-Fock methods, as well as most DFT methods all neglect long-range dispersion. In attempt to model dispersion interactions at reduced computational expense, one approach is to add an empirical potential to one of the quantum mechanical techniques. This approach has been successfully used to model a large variety of systems that involve, or are governed by, dispersion interactions. The accuracy of empirically-corrected density functional theory (DFT-D) and empirically-corrected semiempirical (SE-D) methods are reviewed here. The analysis considers both the ability to reproduce benchmark energies and geometries.