This paper is an analysis of two models on the origin of coercivity in certain crystals. The first model, based upon the work of Kersten. suggested that coercivity in soft magnetic materials is due to the variation in domain wall energy throughout the crystal. The second model attributes coercivity to magnetostatic interactions between domain walls and crystal defects. Recent work on calcium and germanium substituted YIG has shown a defect structure similar to the one assumed in both of these models. Introducing the experimentally observed parameters into the models yields calculated values for the coercivity that are very close to the observed value for the magnetostatic model and very poor for the wall energy variation model.