Raman optical activity has proven to be a powerful tool for probing the geometry of small organic and biomolecules. It has therefore been expected that the same mechanisms responsible for surface-enhanced Raman scattering may allow for similar enhancements in surface-enhanced Raman optical activity (SEROA). However, SEROA has proved to be an experimental challenge and mirror-image SEROA spectra of enantiomers have so far not been measured. There exists a handful of theories to simulate SEROA, all of which treat the perturbed molecule as a point-dipole object. To go beyond these approximations, we present two new methods to simulate SEROA: the first is a dressed-tensors model that treats the molecule as a point-dipole and point-quadrupole object; the second method is the discrete interaction model/quantum mechanical (DIM/QM) model, which considers the entire charge density of the molecule. We show that although the first method is acceptable for small molecules, it fails for a medium-sized one such as 2-bromohexahelicene. We also show that the SEROA mode intensities and signs are highly sensitive to the nature of the local electric field and gradient, the orientation of the molecule, and the surface plasmon frequency width. Our findings give some insight into why experimental SEROA, and in particular observing mirror-image SEROA for enantiomers, has been difficult.