Experimentally determined backscatter factors in mammography can contain significant systematic errors due to the energy response, dimensions, and location of the dosimeter used. In this study, the Monte Carlo method was applied to simulate photon-scattering in tissue-equivalent media and to determine backscatter factors without the interference of a detector. The physical processes of measuring backscatter factors with a lithium fluoride thermoluminescent dosimeter (TLD) and an "ideal" tissue-equivalent detector were also simulated. Computed results were compared with the "true" backscatter factors and with measured values reported by other investigators. It was found that the TLD method underestimated backscatter factors in mammography by as much as 10% at high energies with TLD-100 (3.2- x 3.2- x 0.89-mm3) crystals.