A modified version of this article has been published in the May 2007 issue of the journal Ecological Economics. This paper presents the first comprehensive analysis of municipal-level open space referenda, both the decision to hold a referendum and referendum results, across the entire contiguous United States. We find that municipalities populated by more educated, environmentally aware and politically left-leaning people and that experienced substantial population growth in surrounding areas were more likely to hold open space referenda than other municipalities. Overall, there are fewer statistically significant relationships in the referenda results estimation than in the selection estimation. Referenda support was significantly affected only by the finance mechanism, unemployment rate, education and region dummy variables. We find limited evidence of selection bias in the estimated referenda results equations.