This study looks for one unintended consequence of large-lot zoning: the possibility that it will increase the amount of land converted from rural land cover to suburban lawn, even as it reduces the number of homes that are built in a community. This is one definition of “urban sprawl.” Few zoning studies consider selection bias in the choice of the zoning treatment. When the dependent variable measures parcel development (0 or 1) or land cover change (% of land converted), controls for selection bias are even less common, because these models employ nonlinear link functions. The present study uses inverse propensity score weighting to control for selection bias across six zoning classes in a study of the percentage of land cover change in northwestern New Jersey from 1995 to 2002