The Impact of Voter Evaluations of Leaders’ Traits on Voting Behaviour: Evidence from Seven European Countries

Abstract

This article aims to examine how the evaluation of party leaders’ traits influences voting behaviour. The work incorporates data from election studies across seven countries with different social contexts (Portugal, Spain, Ireland, Germany, United Kingdom, Italy and Hungary). Characteristics of leaders were placed into two groups: competence and warmth, in accordance with the stereotype content model and relevant studies on evaluations of leaders’ traits. Since different personality traits were evaluated within each country, the analysis was based on the two aforementioned dimensions and the political ideology of the leader. Multiple sequential binary logistic regression models were performed to analyse the predictive power of competence and warmth on voting behaviour, controlling for socio-demographic and political ideology variables of voters. Results reinforce the personalisation of politics theory, showing the relevance of leaders’ warm personality traits. As for the competence dimension, it was proven to be decisive only with regard to left-wing party leaders.

Acknowledgements

This work was supported by the FCT (Fundação para a Ciência e a Tecnologia), under the research project entitled ‘The Personalisation of Politics in the 21st Century – A research project on democratic elections’ (Ref. PTDC/CPJ-CPO/120295/2010).

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

1. This problem is not exclusive to this model and is widely documented in the literature; for a more thorough discussion please refer to Bittner (2011).

2. Portugal 2009 (two studies): ‘From the following parties, please tell me which one do you feel closest to?’ and ‘Which party represents your opinions better?’ Spain 2008: ‘Regardless of the party you voted for last Sunday, for which party do you feel more sympathy or consider closest to your ideas?’ Ireland 2002/2007: ‘Which party do you feel closest to?’ Germany 2009: ‘Many people in Germany are inclined to support a particular party for a longer period of time even if they occasionally vote for another party. What about you? In general terms are you inclined to support a party? And if so, which one?’ UK 2005: ‘Generally speaking, do you think of yourself as Labour, Conservative, Liberal Democrat, or what?’ Italy 2013: ‘Is there any political party or civic movement to which you feel closer? If yes, can you tell me which one?’

3. The logistic regression model classification power revealed an overall hit rate of 83.5 per cent (a 33.3 per cent increase compared to the proportional percentage of correct classification by chance: [(3769/7134)² + (3365/7134)²] × 100 = 50.2 per cent), which represents an improvement over chance index of 66.9 per cent ([(83.5 per cent – 50.2 per cent)/(1–50.2 per cent)]*100). According to this result, the model provided a 66.9 per cent reduction in overall classification error over chance. This improvement was significant at p < 0.001, according to a one proportion test. Correct prediction rates of 81.3 per cent for a right-wing vote (sensitivity) and 85.5 per cent for a left-wing vote (specificity) were found.

Concerning the ROC, the final model presented an area under the curve (AUC) of 0.898, which was significantly higher than 0.5 (p < 0.001). This suggested that the model presents a reasonably predictive ability to classify participants in the left/right voting group.

The optimal cut-off value of 0.394 (determined by the intersection of sensitivity and specificity curves) would balance the sensitivity and specificity of the model (0.838 and 0.837, respectively).