When addressing personalization of politics most studies have centered their attention on whether candidates are valued as more important than their parties by citizens. In this study, we follow a wealth of recent scholarship reports that stress another question: are candidates’ non-political attributes considered in the voting decision? (Adam and Maier, 2010). In order to answer this question, I conceptually develop a classification of which attributes could be political, intermediate and non-political. With this in mind, I conduct a candidate-based conjoint experiment where I propose a multidimensional scenario to respondents. The results provide some striking conclusions: while valence issues are the most valued, non-political attributes receive the same importance as policy proposals, being both generally significant. This fact would encourage further investigation including non-political attributes as they have demonstrated to have an impact in the selection of candidates Additionally, I find policy proposals are irrelevant to politically unsophisticated voters. Conversely, dealigned and non-dealigned voters do not show significant differences in the importance that they give to each attribute. However, there is a need for a more systematic study of the personalization phenomenon confronting the external validity of the survey experiment as well as discussing theoretical and methodological decisions I have adopted in this paper.
