The expansion of charter schools—publicly funded, yet in direct competition with traditional public schools—has emerged as a favored response to poor performance in the education sector. While a large and growing literature has sought to estimate the impact of these schools on student achievement, comparatively little is known about demand for the policy itself. Using election returns from three consecutive referenda on charter schools in Washington State, we weigh the relative importance of school quality, community and school demographics, and partisanship in explaining voter support for greater school choice. We find that low school quality—as measured by standardized tests—is a consistent and modestly strong predictor of support for charters. However, variation in performance between school districts is more predictive of charter support than variation within them. At the local precinct level, school resources, union membership, student heterogeneity, and the Republican vote share are often stronger predictors of charter support than standardized test results.