Come hither or go away? Recognising pre-electoral coalition signals in the news

In this paper, we introduce the task of political coalition signal prediction from text, that is, the task of recognizing from the news coverage leading up to an election the (un)willingness of political parties to form a government coalition. We decompose our problem into two related, but distinct tasks: (i) predicting whether a reported statement from a politician or a journalist refers to a potential coalition and (ii) predicting the polarity of the signal – namely, whether the speaker is in favour of or against the coalition. For this, we explore the benefits of multi-task learning and investigate which setup and task formulation is best suited for each sub-task. We evaluate our approach, based on hand-coded newspaper articles, covering elections in three countries (Ireland, Germany, Austria) and two languages (English, German). Our results show that the multi-task learning approach can further improve results over a strong monolingual transfer learning baseline.

[1]  Jeffrey A. Fine,et al.  Negative Campaigning in the Social Media Age: Attack Advertising on Facebook , 2016 .

[2]  James N. Druckman,et al.  Campaign Communications in U.S. Congressional Elections , 2009, American Political Science Review.

[3]  Simone Paolo Ponzetto,et al.  Topic-Based Agreement and Disagreement in US Electoral Manifestos , 2017, EMNLP.

[4]  Christian Elmelund-Præstekær.  Beyond American negativity: toward a general understanding of the determinants of negative campaigning , 2010, European Political Science Review.

[5]  Simone Paolo Ponzetto,et al.  Policy Preference Detection in Parliamentary Debate Motions , 2019, CoNLL.

[6]  Justin H. Gross,et al.  Twitter Taunts and Tirades: Negative Campaigning in the Age of Trump , 2016, PS: Political Science & Politics.

[7]  R. Lau,et al.  Effectiveness of negative campaigning in U.S. Senate elections , 2002 .

[8]  Iryna Gurevych,et al.  AdapterHub: A Framework for Adapting Transformers , 2020, EMNLP.

[9]  Riza Batista-Navarro,et al.  Sentiment and position-taking analysis of parliamentary debates: a systematic literature review , 2019, Journal of Computational Social Science.

[10]  Josef Ruppenhofer,et al.  Fine-grained Named Entity Annotations for German Biographic Interviews , 2020, LREC.

[11]  Issue ownership as a determinant of negative campaigning , 2011 .

[12]  Ming-Wei Chang,et al.  BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , 2019, NAACL.

[13]  Luca Antiga,et al.  Automatic differentiation in PyTorch , 2017 .

[14]  Qiang Yang,et al.  An Overview of Multi-task Learning , 2018 .

[15]  Sebastian Ruder,et al.  An Overview of Multi-Task Learning in Deep Neural Networks , 2017, ArXiv.

[16]  Armin Scholl,et al.  Journalismus in Deutschland (2005) , 2015 .

[17]  M. Dolezal,et al.  Negative Campaigning and the Logic of Retaliation in Multiparty Competition , 2016, The international journal of press/politics.

[18]  Rich Caruana,et al.  Multitask Learning: A Knowledge-Based Source of Inductive Bias , 1993, ICML.

[19]  Brendan T. O'Connor,et al.  From Tweets to Polls: Linking Text Sentiment to Public Opinion Time Series , 2010, ICWSM.

[20]  Heather K. Evans,et al.  Fear and Loathing on Twitter: Exploring Negative Rhetoric in Tweets During the 2018 Midterm Election , 2019, The Roads to Congress 2018.

[21]  Heather K. Evans,et al.  Twitter Style: An Analysis of How House Candidates Used Twitter in Their 2012 Campaigns , 2014, PS: Political Science & Politics.

[22]  S. Bowler,et al.  Voter expectations of government formation in coalition systems: The importance of the information context , 2021, European Journal of Political Research.

[23]  Justin Grimmer,et al.  A Bayesian Hierarchical Topic Model for Political Texts: Measuring Expressed Agendas in Senate Press Releases , 2010, Political Analysis.

[24]  Lukas F. Stoetzer,et al.  How do coalition signals shape voting behavior? Revealing the mediating role of coalition expectations , 2020 .

[25]  Lukas F. Stoetzer,et al.  Estimating coalition majorities during political campaigns based on pre-election polls , 2019, Journal of Elections, Public Opinion and Parties.

[26]  Sebastian Padó,et al.  Swimming with the Tide? Positional Claim Detection across Political Text Types , 2020, NLPCSS.

[27]  Saif Mohammad,et al.  Obtaining Reliable Human Ratings of Valence, Arousal, and Dominance for 20,000 English Words , 2018, ACL.

[28]  Cees van der Eijk,et al.  Measures of Campaign Negativity: Comparing Approaches and Eliminating Partisan Bias , 2019, The International Journal of Press/Politics.

[29]  Luigi Curini,et al.  Negative Campaigning in No-Cabinet Alternation Systems: Ideological Closeness and Blames of Corruption in Italy and Japan Using Party Manifesto Data , 2011, Japanese Journal of Political Science.

[30]  Lysandre Debut,et al.  HuggingFace's Transformers: State-of-the-art Natural Language Processing , 2019, ArXiv.

[31]  Ramit Sawhney,et al.  GPolS: A Contextual Graph-Based Language Model for Analyzing Parliamentary Debates and Political Cohesion , 2020, COLING.

[32]  Yoshimasa Tsuruoka,et al.  A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks , 2016, EMNLP.

[33]  Justin Grimmer,et al.  Text as Data: The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts , 2013, Political Analysis.

[34]  Goran Glavas,et al.  Computational Analysis of Political Texts: Bridging Research Efforts Across Communities , 2019, ACL.

[35]  David F. Damore.  Candidate Strategy and the Decision to Go Negative , 2002 .

[36]  Wouter van der Brug,et al.  When the gloves come off: Inter-party variation in negative campaigning in Dutch elections, 1981–2010 , 2013 .

[37]  Slava J. Mikhaylov,et al.  Scaling policy preferences from coded political texts , 2011 .

[38]  Lukasz Kaiser,et al.  Attention is All you Need , 2017, NIPS.

[39]  Christian Elmelund-Præstekær.  NEGATIVE CAMPAIGNING IN A MULTIPARTY SYSTEM , 2008 .

[40]  J. Kleinnijenhuis,et al.  Polarization in the Media During an Election Campaign: A Dynamic Network Model Predicting Support and Attack Among Political Actors , 2013 .

[41]  M. Dolezal,et al.  Who will attack the competitors? How political parties resolve strategic and collective action dilemmas in negative campaigning , 2015, Party politics.

[42]  Martin Haselmayer,et al.  Friendly fire? Negative campaigning among coalition partners , 2018, Research & Politics.

[43]  Luigi Curini,et al.  Ideological proximity and valence competition: Negative campaigning through allegation of corruption in the Italian legislative arena from 1946 to 1994 , 2010 .

[44]  Michael F. Meffert,et al.  Weighting Parties and Coalitions: How Coalition Signals Influence Voting Behavior , 2017, The Journal of Politics.

[45]  Sebastian Padó,et al.  Masking Actor Information Leads to Fairer Political Claims Detection , 2020, ACL.

[46]  Haiyan Wang,et al.  quanteda: An R package for the quantitative analysis of textual data , 2018, J. Open Source Softw..

[47]  Shuai Wang,et al.  Deep learning for sentiment analysis: A survey , 2018, WIREs Data Mining Knowl. Discov..