The influence of partisan elites on voters’ attitudes toward contested policy issues is a well documented phenomenon (Boudreau & MacKenzie 2014, Bullock 2011, Carsey & Layman 2006, Ciuk & Yost 2016, Druckman et al. 2013, Leeper & Slothuus 2015). There could be different mechanisms behind this influence. Voters may be rational learners who use party positions as information shortcuts (Chong & Druckman 2007, Petersen et al. 2010, Slothuus 2010, Zaller 1996), or motivated reasoners who process information in a biased way (Bolsen et al. 2014, Kunda 1999, Leeper & Slothuus 2014, Slothuus & Vreese 2010, Taber & Lodge 2006).
This paper presents the findings of a series of survey-experiments that have addressed a third mechanism related to the passions fueled by partisan conflicts. Partisan conflicts increase the salience of political identity which, in turn, increases group affection (Druckman et al. 2013, Leeper & Slothuus 2014). As a result of affection, people self-stereotype and conform to perceived group norms by expressing attitudes similar to the perceived typical opinion in their political community (Han & Wackman 2017, Hogg 2003, Iyengar et al. 2012, Malka & Lelkes 2010).
Several studies have addressed the roles of rational learning and motivated reasoning in elite-driven opinion change, and some of them have emphasized the role of conflicts in increasing salience of political identity. However, less efforts have been made to single out the independent effects of political conflicts on attitude adjustments. I constructed a test of conflict-induced attitude adjustment which excludes rational learning and motivated reasoning as possible sources of opinion shifts.
The major hypothesis is that when it comes to opinions on a politicized question, the increase of salience of political conflicts will lead to attitude adjustment in accordance with perceived group norms. This kind of treatment could lead to polarization of attitudes even if respondents don’t get cues on partisan opinions.
I present the findings of two online survey experiments that were carried out on online quota samples of Hungarian inernet-users in 2016 and 2017-18, respectively (the second experiment is a pre-registered study, see: http://egap.org/registration/3041). Subjects were asked about issues that were high on the political agenda in Hungary at the time of the surveys (the recent refugee crisis and the funding of civil organizations), and also about less politicized policy issues as controls.
In the experiments, the treatments manipulate the situational salience of political identity without providing information about partisan elites’ relevant opinions, or arguments about the issues in focus. In particular, some of the treatments increase the salience of political conflicts simply by presenting questions about issues that are highly politicized but unrelated to the target variables.
Our results are in line with the hypothesis about the direct effects of partisan conflicts and feelings of political identity on attitudes toward contested policy issues. In the treatment groups, polarization of attitudes can be observed as left-wing (right-wing) voters’ attitudes become more similar to stereotypical left-wing (right-wing) positions.