Hedonic models are used to estimate implicit prices for water quality in housing markets. Recent studies aggregate sales across large spatial areas in scaled-up models leading to a concern that these models may overlook regional heterogeneity in water-quality preferences. We estimate scaled-up hedonic models comprised of multiple states and individual states and investigate how observations from subregions can differ. We find that the scaled-up model results are driven by select subregions. The results of this study call into question hedonic models using data for large geographic regions where substantial differences may arise across housing markets. / M.S. / Water quality in lakes impacts the prices people pay for lakefront properties. However, these effects can vary across different housing markets. We study whether owners of lakefront properties throughout the northeast and upper Midwest are willing to pay the same amount for water in lakes. We find that in multi-state housing markets, the effects from one state can dominate the overall results and there is likely heterogeneity in preferences across housing markets.