Political Behaviour/Sociology

F17 - Applied Experimental Methods in Political Behaviour

Date: Jun 5 | Time: 10:15am to 11:45am | Location:

Chair/Président/Présidente : Vince Hopkins (University of British Columbia)

Discussant/Commentateur/Commentatrice : Christian Schimpf (University of British Columbia)

Experimental methods in political science have become increasingly central to understanding how individuals and groups make political decisions. From survey experiments to field experiments, these methods offer powerful tools to test causal relationships and explore political attitudes, participation, and behavior dynamics in real-world contexts. This panel brings together a range of perspectives on how experimental techniques advance our knowledge of political behaviour in Canada and beyond. Jointly, the papers demonstrate how experimental techniques can be used to study questions of great relevance that closely relate to the conference's theme, "The Politics of Belonging." The four papers use different experimental methods to study the importance of immigrants' longstanding partisan identities for political incorporation in host countries (Paper 1), the role of social desirability bias in misinformation reporting (Paper 2), the role of elite cues for attitudes towards and trust in the Bank of Canada (Paper 3), and how Canadians evaluate political candidate authenticity (Paper 4).

Can Conversational AI Influence Party Choice Among Immigrants?: Yamil Velez (Columbia University), Semra Sevi (University of Toronto), Daniel Rubenson (Toronto Metropolitan University), Donald P. Green (Columbia University)
Abstract: Political incorporation can be hindered when immigrants struggle to translate their political preferences across different contexts. This study investigates how immigrants' longstanding political attachments and partisan identities shape their party identification in host countries. While immigrants often retain political ideologies and loyalties from their homelands, these may not align with the political landscape of their new country. However, information about how party systems differ across countries can be difficult to obtain. To address this, we developed a chatbot that uses information about party systems in immigrants' countries of origin and integrates a large language model with retrieval-augmented generation to provide balanced, official party information from the host country. We are currently conducting an experiment in Canada with 2,000 immigrants to assess how effectively this "political translation" bot helps improve their understanding of party positions and fosters new partisan attachments. NOTE - there are six authors in total on this paper. As CPSA only allows four entries, here is the full list of authors: Authors: • Yamil Velez (yrv2004@columbia.edu) -- Assistant Professor, Department of Political Science, Columbia University • Semra Sevi (semra.sevi@utoronto.ca)-- Assistant Professor, Department of Political Science, University of Toronto • Daniel Rubenson (rubenson@torontomu.ca) -- Professor, Department of Politics & Public Administration, Toronto Metropolitan University • Donald P. Green (dpg2110@columbia.edu) – Professor, Department of Political Science, Columbia University • Phil Triadafilopoulos (t.triadafilopoulos@utoronto.ca) – Professor, Department of Political Science, University of Toronto • Vince Hopkins (vince.hopkins@ubc.ca) – Assistant Professor, Department of Political Science, University of British Columbia Presenter: Semra Sevi

Misinformation Correction and Preference Concealment: Evidence from Disaster Relief during Hurricane Helene: Boyoon Lee (Queens University), Ted Hsuan Yun Chen (George Mason University)
Abstract: Surveys that examine respondents' reactions to misinformation correction are commonly used to assess the effectiveness of accurate information in combating misinformation. However, concerns have been raised regarding the validity of these responses due to potential preference concealment — being corrected may signal that the misinformation is socially undesirable, leading respondents to conceal their genuine, but socially disapproved, preferences. We use a list experiment to assess the impact of social desirability bias on survey responses to misinformation correction, and how this varies across partisan groups. Specifically, our empirical test asks respondents to read a news article about Hurricane Helene, during which Trump made false statements about disaster relief funds being diverted toward migrants who entered the country illegally, and was called out by political elites across the aisle. We experimentally vary the information across two dimensions: the correction of misinformation (misinformation alone v.s. misinformation and correction) and the source of the correction (own party v.s. opposite party). The Hurricane Helene case has important features for our study as climate disaster relief has been a relatively non-partisan issue in the U.S., making Trump's claims a relative shock to the public opinion system, and making corrections from both Democrat and Republican sources credible treatments. This study provides new insights into the role of partisan motivations in concealing sincere beliefs and opinions in the face of misinformation.

The Impact of Priming Cues on Public Attitudes: The Case of Central Bank Digital Currencies: Anwar Sheluchin (McMaster University), Jordan Mansell (McMaster University), Ori Freiman (McMaster University), John McAndrews (McMaster University)
Abstract: This paper investigates whether elite cues shape public attitudes toward central bank digital currencies (CBDCs) and influence trust in the Bank of Canada. Amid growing concerns over central bank independence and the politicization of monetary policy, understanding how elite messages affect public perceptions of central bank legitimacy is crucial. The political behaviour literature emphasizes the role of elite cues in shaping public opinion, particularly in policy areas where citizens have limited information and rely on cues from trusted sources. This study contributes to that literature by hypothesizing that exposure to positive or negative elite cues can significantly impact trust in the Bank of Canada. To test this hypothesis, we employed a within-subjects experimental design. Our experiment involved priming participants with varying elite statements about CBDCs, including statements from specific political elites—one endorsing cryptocurrency and another opposing CBDCs. The findings suggest that individuals exposed to a positive prime or to a political elite's endorsement of Bitcoin were subsequently less trusting of the Bank of Canada's ability to issue a secure CBDC. However, exposure to a negative prime or a political elite's opposition to CBDCs did not significantly impact trust in the Bank. NOTE: The CPSA conference submission feature only allows for up to four authors. There are five authors on this paper. Here is the full list: Authors: • Anwar Sheluchin (mohaa33@mcmaster.ca) -- PhD Student, Department of Political Science, McMaster University *Lead and presenting author • Jordan Mansell (mansellj@mcmaster.ca) -- Postdoctoral fellow, Digital Society Lab, McMaster University • Ori Freiman (freimano@mcmaster.ca) -- Postdoctoral fellow, Digital Society Lab, McMaster University • John McAndrews (john.mcandrews@mcmaster.ca) -- Adjunct assistant professor, Department of Political Science, McMaster University • Clifton van der Linden (cliff.vanderlinden@mcmaster.ca) -- Associate professor, Department of Political Science, McMaster University Presenter: Anwar Sheluchin

Who's "Really" Authentic? The Role of Race, Gender, and Party in Shaping Perceptions of Candidate Authenticity: Emil Francis (Memorial University of Newfoundland), Amanda Bittner (Memorial University of Newfoundland)
Abstract: In 2019, Stiers et al. advanced a measure of authenticity based on a six-item battery, arguing that voters differentiate between candidates based on perceptions of authenticity, which can be distinguished from other personality traits. Their work suggests that perceived candidate authenticity is consequential for voters' evaluations of parties, leaders, and vote intention. We build on this work by assessing the extent to which Canadians evaluate candidate authenticity, extending the analysis to consider the role of other key demographic variables in influencing perceptions of representatives. Using an intersectional approach, we analyze the results of a survey experiment embedded in an original survey fielded in Canada in the spring of 2024. We look at the extent to which race, gender, and party of representatives intersect to influence voters' perceptions of their authenticity.