The Decision Science of Voting: Behavioral Evidence of Factors in Candidate Valuation

Despite decision science have increased our understanding of human decision-making in different contexts, voters’ decision has been studied less from this point of view. Therefore, we investigated, how electorate- and candidate-related factors affect electorate’s (N=1334) valuation to the Prime Minister candidates (N=11) on the multiparty democracy. Electorates valuated candidates individually and through pairwise candidate comparison. We collected the data by using anonymous questionnaire and sent it via mass emailing and social media. We applied linear mixed-effects and Bayesian network models to analyze the data. Electorate-related variable Valence and candidate-related variables Trustworthiness and Righteousness was found as the strongest main effects. The pairwise analysis comparison highlighted voters’ personal characteristic. In particular, the interactions associated to valence, arousal and gender had high effect only in pairwise comparisons. Our results suggest that the pairwise comparisons - which is typical for elections, e.g., in USA - highlights the importance of emotional and gender-related factors