Theories of dynamic representation (that is, the linkage between changes in aggregate public opinion and government response over time) are well established. But they are based upon global measures of both public opinion and policy responsiveness. Recent studies have shown the value of considering policy responsiveness separately in specific issue-domains. This study tests more broadly theories of dynamic representation on a number of individual policy areas.
The investigators will produce an innovative database allowing the systematic integration of public opinion measures with existing measures of public policy priorities. The project allows for the first time the large-scale integration of public opinion and public policy measures across a multitude of policy domains.
Few questions are as central to the notion of representative democracy as the linkage between public opinion and public policy response, so this project will have important contributions to make on a core issue of American government. It also contributes to the infrastructure available to all scholars. In addition, the project provides training for graduate and undergraduate students and provides the integrations of the new data resources into teaching a number of different levels.
Creation of Policy-Specific Mood Series The essential promise of the proposal was the creation of as many time series of policy-specific mood as the data richness permitted. That has been accomplished. We have created in fact 59 such series corresponding to either specific policy codes of the Policy Agendas Project or to master category codes in addition to three series that are politically important but do not correspond to particular codes. Integration into the Policy Agendas Web Site As we have also promised, our created data series are available for free access and scholarly use from the Agendas Project database (www.policyagendas.org) (residing at the University of Texas at Austin). Scholars may easily download the entire set of 62 series along with accompanying documentation on series creation. Alternatively the individual series may be accessed one at a time with the site's common analysis tool, which allows users to mix series from the various parts of the Agendas Project with one another for graphic display. This permits, for example, graphic display of a policy-specific mood time series with other indicators (e.g., congressional hearings activity) for the same policy code. The analysis tool was designed for ease of access and use, permitting extensive use in the undergraduate classroom environment. A New Mood Creation Tool Not anticipated by the proposal, we have also created a capability for users to create their own series to their own particular needs. One of the early feedbacks from public presentation of our work was from policy specialists who did not wish to be bound by our (and Agendas Project) definitions of policy areas. An education specialist, for example, complained that much of the survey data on his topic concerned federal spending on education, a topic quite different from his focus on school vouchers and various aspects of the education reform agenda. From that feedback we envisioned a tool that would allow users to see the kinds of survey items that fell into one or multiple topic codes and then make a selection of the particular sorts of issues that fell into their definition of the policy domain. The Dyad Ratios algorithm that underlies the creation of mood estimates was thus built into the Agendas website which now has the capability of generating user-determined definitions of policy mood and creating them in real time for download and use. One of the challenges of this task was making it possible to use the public opinion database without making it possible to observe the individual survey-level data, which would have violated re-dissemination agreements of the data providers. As implemented, the system generates time series of user-defined policies while still making it impossible to observe or estimate the original "top line" data.
