Practical Political Knowledge
This document summarizes my work within the ELECDEM network. I argue that politically uninformed people are less likely to provide information on their political attitudes but they are also more prone to have economically left-wing ideological leanings. This leads to the systematic overrepresentation of right-wing attitudes in surveys and may also be harmful to the interests of those who would benefit most from communitarian policies. Using linear and Poisson regression techniques on data from an experimental survey carried out in September 2009 in Hungary, I show that aggregate changes in political attitudes are likely to be revealed when correcting for item non-response missing data for attitudinal variables.
Notwithstanding the relationship between political knowledge and opinionation, I argue that political knowledge has an inevitable ideological bias generated by the distribution of right wing and left wing arguments in the media and printed press. I find no conclusive evidence for this claim on cross-national data; however, null and substantively insignificant results are often found in the literature on political literacy, suggesting that it is either the theory of information effects that is wrong, or we fail at making reliable measurements of political knowledge. In order to shed light upon this matter, I investigate the sensitivity of the reliability of political knowledge scales to variations at the metric and conceptual level. We find that the format of the questions that are used for measuring political knowledge accounts for up to one third of the variance of reliability on data from the second module of the Comparative Study of Electoral Systems. The bootstrapped estimates show that this result is robust, and the same pattern holds for various measures of reliability and convergent validity.
On Dutch Vote Compass data, we show that there may be issues of conceptualization as well that affect the results cited in the literature on information effects. We find that the classic triadic model of knowledge acquisition (capability-opportunity-motivation) needs to be revised to correctly specify capability as a moderator rather than a direct factor influencing political knowledge. Finally, we find with a correlational, incomplete confirmatory factor analysis, that the concept of political knowledge is not unidimensional, as the literature assumed to date. All these findings point to the conclusion that the utility of political literacy is likely to be understated in the empirical literature. Some of the null findings and weak relationships found thus far between political knowledge and indicators of participation, turnout, opinionation, tolerance or pro-democratic attitudes, are in fact stronger in reality than empirical results would suggest.
Download the full report: WP09_Paul Weith_ESR Final Report (PDF 2,886KB)