We compare empirical and theoretical models of polar cap boundary locations and ionospheric convection patterns for the Geospace Environment Modeling (GEM) Boundary Layer Campaign Interval during January 1992. The empirical model is described by Lyons [this issue>, and the theoretical model is described by Toffoletto and Hill [1993>. The empirical and theoretical models agree on the large-scale pattern of polar cap flow, including its dawn-dusk deflection in response to the dawn-dusk component of the interplanetary magnetic field (IMF). There is, however, a discrepancy with respect to the magnitude of the flow, as measured by the polar cap potential drop. This parameter is an input to the theoretical model, for a given solar wind/IMF parameters, based on previous statistical studies, and the standard (empirical) input exceeds the empirical value obtained for these cases by a factor of order 2. There is also a discrepancy with respect to the size of the polar cap, which is smaller in the theoretical model than in the empirical one. The size discrepancy is partly, but not entirely attributable to the potential-drop discrepancy. The comparison provides clear directions for improvement of the theoretical model and perhaps also for refinement of the empirical modeling technique. ¿ 1998 American Geophysical Union