A comparison of how well three different electric field models can predict the storm time plasmapause shape is conducted. The magnetic storm of 17 April 2002 is selected for this event, and plasmapause locations are extracted from images from the EUV instrument on the Imager for Magnetopause-to-Aurora Geomagnetic Effects (IMAGE) satellite throughout the main phase and recovery phase of the event. The three electric field descriptions are as follows: the modified McIlwain E5D analytical formula, the Weimer statistical compilation from low-Earth orbit satellite data, and a self-consistent Poisson equation solution for the subauroral potential pattern. It is found that all of the models have certain strengths and weaknesses in predicting the plasmapause location during this storm. The modified McIlwain model did well on the nightside but not on the dayside because the electric fields near noon are too small (analogous to too large of a conductance in the subauroral dayside ionosphere). The Weimer model did well overall, but the resulting plasmapause is usually smaller than the observed one because the electric fields are a bit too strong in the inner magnetosphere (perhaps because of an ionosphere-magnetosphere mapping problem). The self-consistent model is also quite good in general, but the plasmapause in the postmidnight sector was always inward of the observed one. This is because of too low a conductance at the location of the field-aligned currents that close the partial ring current. It is concluded that the latter two models provide a sufficient description of the storm time development of the plasmaspheric morphology during this storm, with the self-consistent model being the best choice. Another conclusion is that plasmapause locations extracted from EUV images should be compared with peak density gradients from model results rather than with any one isocontour of the cold plasma density itself.