Do democratic systems provide a meaningful role for ordinary citizens to influence public policy? To what extent does the answer depend on political institutions, shared responsibility among governments for the policy in question, and the particular issue? Answers to these questions are essential to gaining a clear understanding of democratic performance and the critical issues of institutional design that can make democracies function better.
This project provides substantive answers to both questions at the same time that it develops innovative tools to address the questions. The intellectual merit of this project thus has two components. First, it sheds light on why policies written at one level of government are not always faithfully implemented by lower levels of government. This provides both basic knowledge about the functioning of democratic institutions and practical knowledge that can be used to design more efficient and effective public policies. The second element of the project's intellectual merit rests on the fact that progress in the "policy responsiveness" field of political science has depended on increasingly sophisticated methodologies for estimating public opinion for small areas (such as states and localities in the US) or small time periods, so that it has benefitted from advances in the area of applied statistics known as "small area estimation." This project contributes to the science of small area estimation by extending the current best practice (Multilevel regression with post-stratification, or "MRP") with an innovative alternative estimation procedure that treats the challenge of small area estimation as a missing data problem.
The project also makes several broader contributions. Most broadly, the project expands understanding of the workings of democratic institutions and yields findings of interest to scholars, policy makers, and citizens alike. In addition, as an end product of the project's small area estimation efforts, its project creates and makes widely available a dataset of current estimates of public opinion for states, counties and larger municipalities for multiple policy domains. This will be useful for academic, government and private sector researchers in many fields.
