CEU Electronic Theses and Dissertations, 2016
| Author | Weith, Paul Thomas |
|---|---|
| Title | Information Effects Across the Globe: Does Political Knowledge Matter, and How Do We Know It? |
| Summary | The study of the various merits of citizens' political knowledge has a rich tradition in political science research, by virtue of its purported pivotal role in maintaining democratic accountability. How knowledge gains can influence politically relevant outcomes is a frequent question in the study of citizens’ political behavior, and the answers vary considerably across political contexts and across studies. This thesis offers an in-depth investigation of the origins of this variation in electoral democracies spanning the globe. In doing so, it builds on insights from an eclectic range of research fields from survey methodology and measurement theory, to cognitive psychology and comparative politics. First, I show that a selection of technical properties of questionnaire items can account for more than half the variation in the reliability of knowledge measurements. I demonstrate that this can drastically affect the ability of political knowledge to explain outcomes of interest for political scientists. However, I find that the perils of measurement error can be circumvented by modeling techniques previously proposed for estimating information effects on vote choice (Bartels, 1996; and Delli Carpini and Keeter, 1996). The outcome of interest is regressed on political knowledge, a set of demographic variables and all the possible interactions between knowledge and the other variables in the model. By subtracting the expected value of the outcome variable from the same statistic under conditions of maximum political knowledge for every respondent, a total effect of political knowledge can be estimated. As it was previously suggested that most of the effect of knowledge on political preferences occurs via indirect causal pathways (Zaller, 1992), interactive models are expected to better approximate the true effects of knowledge, compared to models that only estimate a direct effect. This claim, however, has not been tested prior to this dissertation. The empirical tests in the final substantive chapters of this thesis compare the two modeling approaches: the complex model with multiple interaction terms and the simple model with only a direct effect of knowledge included. In Chapter 3 I argue that the impact of knowledge on vote choice should be contingent upon the simplicity of the decision task (Lau et al., 2008). A meaningful decision in an election where there are few and easily distinguishable political alternatives does not require any amounts of expertise, thus information effects are expected to be weak. In Chapter 4, I argue that people develop positive attitudes towards the political system either by learning its virtues, or by having an affective attachment to their political community. In small, unitary countries, where there are few and easily distinguishable political parties, little knowledge is necessary for citizens to develop positive attitudes towards the system. I find more support for all these theories using the more complex interactive model, but the direct-effect model performs remarkably well in most cases. This thesis offers ample evidence of information effects on political behaviors and attitudes in a wide range of electoral democracies across the globe. It corroborates findings previously reported in the literature, and demonstrates the ability of political institutions to mitigate information effects on vote choice, turnout, political efficacy and satisfaction with democracy. It also challenges the highly influential assertion that political knowledge has a primarily indirect effect on political preferences. |
| Supervisor | Tóka, Gábor |
| Department | Political Science PhD |
| Full text | https://www.etd.ceu.edu/2016/weith_paul.pdf |
Visit the CEU Library.
© 2007-2021, Central European University