Replication Data for: Representation in Municipal Governments: A Replication and Extension to School Districts

In "Representation and Municipal Government," Tausanovitch and Warshaw (2014) examine the extent to which cities in the U.S. are responsive to the views of their citizens. Using a multilevel regression and poststratification (MRP) model, the authors pool survey data to predict citizen preferences for U.S. cities with more than 20,000 citizens. They then compare citizen policy preferences with policy outputs at the municipal-level. The authors find that policy outputs correspond with the liberal-conservative policy preferences of citizens, but that institutions (e.g. having an elected mayor, partisan elections, etc.) have little to no impact on policy responsiveness. We provide two extensions. First, we replicate their core findings and use their estimated ideology scores to show that school district policies are strongly associated with citizen-level policy views. Second, we follow their entire methodological process with new survey data and create true school district-level estimates of ideology. These initial results suggest that school district policies largely correspond with the views of the citizens represented by them.

在"Representation and Municipal Government"一文中，陶萨诺维奇（Tausanovitch）与沃肖（Warshaw，2014）探究了美国城市对其民众观点的响应程度。研究团队采用多层回归与事后分层（multilevel regression and poststratification，MRP）模型，汇集多份调查数据，对人口规模超2万的美国城市的民众政策偏好进行预测。随后他们将民众的政策偏好与市政层面的政策产出展开对比分析。研究发现，市政政策产出与民众的自由派-保守派政策偏好高度契合，但制度因素（如民选市长、党派选举等）对政策响应度几乎无影响。本研究做出两点拓展：其一，我们复现了其核心研究结论，并借助他们测算的意识形态得分，证实学区政策与民众层面的政策观点存在显著关联；其二，我们采用全新调查数据完整复刻了其研究方法流程，生成了准确的学区层面意识形态测算值。初步研究结果表明，学区政策大体上与其所代表的民众观点相契合。