Understanding party politics using feeling barometers and generalized unfolding models

Jour Fixe talk by Konstantin Käppner on January 31, 2018

Konstantin Käppner was a Doctoral Fellow in 2017 and is affiliated with the Department of Politics and Public Administration. 

Political Scientists, journalists and pundits alike often employ spatial models and metaphors to describe political competition among parties and candidates. Far-Right-Wing parties are for instance pitted against centrist or leftists parties and voters are said to choose among parties based on how "close" they are to them position-wise on relevant issues. This abstraction from specific issues to a spatial represtentation has been very influential in developing models of voting, coalition-formation and party strategies, but has not been without criticism. First, many relevant issues that matter to voters cannot be subsumed under a spatial metaphor, but rather follow a distinct, non-spatial logic. Second, spatial and non-spatial factors and the predictions associated with them are hard to test empirically due to the great requirements they put on survey data and the lack of an accepted measurement standard. 

In this talk, Konstantin drew on joint work with Susumu Shikano to show that applying extended unfolding models from psychology on widely available feeling thermometers ratings alone can be used as a widely applicable measurement of party positions and non-spatial components. Konstantin illustrated the utility of the proposed method using different data sources from the 2017 German Federal election and gave an overview over the German 7-party system and its evolution since 1976.