Observations of novae at radio frequencies provide us with a measure of the total ejected mass, density profile, and kinetic energy of a nova eruption. The radio emission is typically well characterized by the free-free emission process. Most models to date have assumed spherical symmetry for the eruption, although for as long as there have been radio observations of these systems, it has been known that spherical eruptions are too simplistic a geometry. In this paper, we build bipolar models of the nova eruption, assuming the free-free process, and show the effects of varying different parameters on the radio light curves. The parameters considered include the ratio of the minorto major-axis, the inclination angle, and shell thickness. We also show the uncertainty introduced when fitting spherical-model synthetic light curves to bipolar-model synthetic light curves. We find that the optically thick phase rises with the same power law (S ν ∝ t 2 ) for both the spherical and bipolar models. In the bipolar case, there is a "plateau" phase-depending on the thickness of the shell as well as the ratio of the minor-to major-axis-before the final decline, which follows the same power law (S ν ∝ t −3 ) as in the spherical case. Finally, fitting spherical models to the bipolar-model synthetic light curves requires, in the worst-case scenario, doubling the ejected mass, more than halving the electron temperature, and reducing the shell thickness by nearly a factor of 10. This implies that in some systems we have been over-predicting the ejected masses and under-predicting the electron temperature of the ejecta.