Pre-Electoral Coalition Signals
Final Report Abstract
Under which conditions are parties willing to send coalition signals during election campaigns? To address the overarching research question, the project focused on two aspects. First, it was crucial to understand how pre-electoral coalition signals influence voting behavior. Second, we investigated under which conditions parties are willing to send coalition signals. To study how coalition signals shape voting behavior, we collected different types of data. We conducted different survey experiments in the United Kingdom (2020) and in the contexts of the general elections in Sweden (2018), Ireland (2020), New Zealand (2020), and Germany (2021). To test when parties signal their preferred coalitions, we set up a cross-country database of electoral coalitions in 398 legislative elections in 22 advanced industrialized democratic countries from 1946 to 2014. We studied press releases sent by German parties (2002-2005), and tweets sent by parties and MPs in Germany (2017-2020) and Spain (2018- 2020). We collected coalition signals from newspaper articles in 17 elections in five countries. Using this dataset, we collaborated with computer scientists from the University of Mannheim to train a classifier that automatically detects coalition signals from newspaper articles. Our results provide several insights for the influence of coalition signals on voting. First, coalition signals affect voting by changing voters' expectations about which coalitions are likely to form. Second, voters prefer a clear link between their party vote and the expected coalition government. Third, breaking coalition promises reduces voters’ propensity to vote for such parties. Fourth, motivation, information, and capabilities are preconditions for strategic voting. Regarding parties’ coalition signaling behavior, we find that parties consider the ideological fit between themselves and their potential coalition partners. We find a tendency for parties to form pre-electoral coalitions with parties on the same side of the ideological spectrum, that are ideologically close, and that are in the same ideological bloc in the Spanish case. Parties primarily send positive signals to parties that would be a plausible coalition partner. Our research has been featured by newspapers such as The National (https://www.thenationalnews.com/world/europe/sweden-to-face-political-uncertainty-as-far-right-makes-electoral-gains-1.768838) and Sueddeutsche Zeitung (https://www.sueddeutsche.de/politik/bundestagswahl-prognose-vorhersage-1.5385101?utm_source=Twitter&utm_medium=twitterbot&utm_campaign=1.5385101), as well as by the project zweitstimme.org.
Publications
- (2020). Coalition policy perceptions. The Journal of Politics, 82(4), 1458-1473
Bowler, S., Gschwend, T., & Indridason, I. H.
(See online at https://doi.org/10.1086/708629) - (2020). Estimating coalition majorities during political campaigns based on pre-election polls. Journal of Elections, Public Opinion and Parties, 30(1), 126-137
Stoetzer, L. F., & Orlowski, M.
(See online at https://doi.org/10.1080/17457289.2019.1582533) - (2020). How do coalition signals shape voting behavior? Revealing the mediating role of coalition expectations. Electoral Studies, 66, 102166
Bahnsen, O., Gschwend, T., & Stoetzer, L. F.
(See online at https://doi.org/10.1016/j.electstud.2020.102166) - (2021, November). Come hither or go away? Recognising preelectoral coalition signals in the news. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 7798-7810
Rehbein, I., Ponzetto, S. P., Adendorf, A., Bahnsen, O., Stoetzer, L. F. & Stuckenschmidt, H.
(See online at https://doi.org/10.18653/v1/2021.emnlp-main.615)
