The supposed personalization of voting behaviour is broadly discussed in election studies. Based on trends regarding the development of society, media and the political system, it is argued that people focus more and more on candidates, especially front runners, than on party identifications and issues. The empirical results for Germany are ambiguous, though. Besides the fact that different periods of time are regarded, the ambiguity might be due to the fact that contextual factors are often ignored.
The paper argues that vote-specific factors like the constellation of candidates, especially differences in the voters’ estimation of the candidates, may influence the importance of candidate evaluations for voting behaviour. Hence controlling for these factors is necessary when analysing the personalization thesis. In the paper results of a multilevel-analysis, containing individual and election specific data shall be presented. The analysis draws on data from the German state elections, reaching back to the 1960s.
