Government Performance, Valence Judgments, and The Dynamics of Party Support

University Of Texas At Dallas, Richardson TX

Investigators

Abstract

This research project advances knowledge of the dynamics and determinants of party support in democratic polities. One aspect of the research concerns dynamic relationships between public service delivery and aggregate party support. Acutely aware of the importance of these relationships, governing parties of varying ideological hues have made repeated attempts not only to bolster the supply of public services, but also to curb the electorate's demand for them. The latter efforts have met with relatively little success. Many voters remain convinced that government can, and should, be heavily involved in fields such as education, health, welfare, and transportation, that privatization schemes often go awry, and that government is responsible when public services go badly or are not delivered at all. To date, however, research on factors affecting the aggregate dynamics of party support has focused heavily, indeed almost exclusively, on the impact of economic conditions and various "one-off" political events. By restricting attention to economic conditions and salient events, existing studies have ignored the consequences of public reactions to government performance in a wide variety of highly salient policy areas such as health care, education, and national security. Building on the results of earlier research on the determinants of party support conducted by the principal investigator and colleagues, the study will specifies and tests rival models of how government performance evaluations in several important policy areas affect the dynamics of aggregate-level support for governing and opposition parties in a major mature democracy, Great Britain. A second aspect of the research concerns modeling the individual-level dynamics of party support. Here, the principal focus is gathering the large-N, multi-wave panel data needed to address longstanding controversies regarding the individual-level dynamics of partisan attachments. Although these controversies have major theoretical implications for understanding the nature of party support in democracies new and old alike, their resolution has been hindered by the lack of sufficiently large multi-wave panels, disputes regarding the proper measurement of partisan attachments, and the absence of suitable statistical tools for analyzing the latent-level dynamics of nominal-scale measures of partisanship. Such tools are now available, and the research gathers the requisite large N panel data needed to analyze individual-level latent variable dynamics of partisan attachments measured using both the traditional British Election Study (BES) party identification question sequence and an alternative party supporter battery. The aggregate-level time series and individual-level panel data needed to conduct the research are gathered in 48 consecutive monthly surveys of the British electorate. Surveys are be conducted by YouGov, Britain's premier internet survey firm. Each monthly survey is conducted with a national sample of 1,000 persons 18 years of age or older, yielding a total N of 48,000 cases. Multi-wave national panels, including a four-wave panel with an N of approximately 5,000 cases will be embedded in the surveys. The large size of the cross-sectional and panel surveys facilitate theoretically interesting sub-group analyses. Aggregate time series analyses will be informed by recent work by the principal investigator and colleagues on the specification and testing of rival fractional error correction models. Individual- and aggregate-level data gathered in the study will be deposited annually with the ICPSR Data Archive, and also will be available for downloading from the project website. The website also will provide monthly "DataCam" updates of the evolution of time series variables. Project findings will be presented at major scholarly conferences, and papers and technical reports will be available for downloading from the website.