F12 - Workshop: Experiments in Political Science 2
Date: Jun 3 | Time: 02:00pm to 03:30pm | Location:
Chair/Président/Présidente : Jordan Mansel (Université du Québec à Montréal)
Discussant/Commentateur/Commentatrice : Valerie-Anne Mahéo (Université de Montréal)
Cooperation and Defection in an Iterated Prisoner’s Dilemma: Do Liberals and Conservatives Display Differences in Social Cognition?: Jordan Mansell (Université du Québec à Montréal), Michael Bang Petersen (Aarhus University)
Abstract: In its most simple form, the research question that we want to answer is: Do individuals with different political values differ in their general social-cognitive strategies? Research into ideological differences links liberal and conservative ideological positions to variation in cognition and decision-making in response to environmental stimuli including risk, ambiguity, and uncertainty. We apply an evolutionary framework and investigate whether the differences in cognition and decision-making observed in liberals and conservatives reflect alternative adaptive strategies for social interactions? We recruit a sample of (N=450) liberals and conservatives to participate in multiple rounds of an iterated prisoner’s dilemma. Using a multi-wave approach to data collection we obtain a sample of liberals and conservatives which is matched on age, education, income, and strength of ideological orientation. We hypothesize that both ideological groups will converge to cooperative behaviour however, cooperation in conservative participants will be more sensitive to social defections. On multiple measures of ideological orientation, we find strong evidence liberals and conservatives significantly differ in their: (1) initial willingness to cooperate; (2) overall cooperation; and, (3) forgiveness of a social defection. We implement our study online in real-time using the OTree tool for experimental research on a Heroku server.
Framing Public Opinion about the Brazilian Social Security Reform in 2019: An Analysis of a Three-round Experiment: Wladimir Gramacho (University of Brasilia)
Abstract: This paper describes dynamic changes on how Brazilians responded to two frames while assessing the social security reform debated in the country in 2019. One of the frames - the loss frame - explained the reform proposed by the government as a means "to reduce public expenses with pensions". This frame was dominant in media coverage and also in public communications of government officials and politicians. The other one - the gain frame - explained the reform as a means "to increase public expenses in health and education". This frame was only incidentally used in public debate. Our main hypothesis is that extensive repetition of a frame will reduce or even cancel out its effect on the opinion of individuals about the reform, while when it is not used, it's potential effect on opinions should be stable overtime. In order to test our hypothesis, we conducted three survey experiments: one in November 2018 (by telephone), before the inauguration of the new President; another in January 2019 (online), weeks after the announcement of the reform as a priority for the government; and the last one in March 2019 (also online), when the reform had been sent to Congress. The results mainly confirm our expectation, as the effect of the loss frame disappears and the influence of the gain frame keeps steady. Nevertheless, different patterns appear in the behavior of different social subgroups, separated by gender, age and opinion about the government.
Not so Normatively Appealing Citizen: Does Party Identification Trump Issue Opinions in Multiparty Systems?: Semih Cakir (Université de Montréal)
Abstract: Democratic theories argue that citizens influence the decision-making process by electing representatives to act on their behalf in electoral democracies. Citizens are also considered to be relatively well informed about issues and party stances on these issues. This way, they can serve as a check on politicians by holding them accountable for policies they promise to implement and for their past performance. However, evidence against this normative democratic citizenship is growing. Party identification is shown to blur this picture of democratic citizenship. Far from holding politicians to account for deviating from their preferences, voters indeed align their issue preferences to fit the positions of their leader/party even when they contradict their initial position on the issue. These studies, nonetheless, are mostly limited to the bipartisan American context, where change in party position is less likely to produce a party conversion to a polar opposite party. Thus, citizens are mechanically encouraged to follow their party in such a context. However, we know less whether this relation holds in multiparty contexts. Does the presence of other parties that share individual’s initial opinion prevent her from converting her opinion to fit that of her party and lead her to change her party identification? In this article, I proceed as follows. First, I establish the causal direction by manipulating party positions in an experimental study in Canada, the United Kingdom and Germany. Second, I test my findings with panel data in these countries.
How Norms Shape the Nature and Origins of Mass Belief Systems: Mark Pickup (Simon Fraser University), Eric Groenendyk (University of Memphis), Erik Kimbrough (Chapman University)
Abstract: Are Americans’ policy preferences constrained by ideological belief systems? If so, what does this imply? After more than fifty years of research, consensus remains elusive. We bring clarity to this debate by addressing three vital questions: First, researchers operationalize constraint based on their own expectations, but is there really normative consensus among ideology group members on “what goes with what?” Using an incentivized coordination game, we find substantial variability in the clarity of ideological norms across issues. Second, the extant literature equates lack of constraint with political ignorance, but how many seemingly “ideologically innocent” voters know “what goes with what” yet choose to flout these norms? When we measure political beliefs and knowledge of ideological norms separately, it becomes clear that ignorance and pragmatism are typically conflated. Third, does constraint facilitate accountability or does it just represent strong adherence to ideological norms? Using a survey experiment, we find that priming norms increases adherence. This suggests that the most ideologically constrained are the strongest norm followers and perhaps not best suited to ensure accountability.
Experiments in Child Care Policy: Does Information Make Parents 'Better' Consumers of Care?: Samantha Burns (University of Toronto), Adrienne Davidson (Queen's University), Linda White (University of Toronto), Michal Perlman (University of Toronto)
Abstract: Does receiving information about the licensing of child care options change parental preferences for child care? While we might expect parents to update their preferences towards licensed child care upon learning about the lack of oversight in some care arrangements, this outcome is by no means guaranteed. Experiments can nudge participants to be more accurate in their decision making by informing them of the importance of a task can be motivated to engage in more complex decision making (Kunda, 2000, p. 481). However, information does not ensure a shift prior attitudes in citizens (Nyhan & Reifer, 2010) or political officials (Baekgaard et al., 2017). Child care decisions are highly complex, shaped by a combination of structural limitations (such as cost and location), informational barriers, as well as ideas about what the goal of care should be (e.g. education? an environment that feels like home?). Drawing on a conjoint survey of parents in the City of Toronto, this paper examines the impact on information for parental preferences for child care. Parents were randomly assigned into an information treatment about the three types of legal child care in the province of Ontario, including information about the degree of regulatory oversight (which ranges from no functional oversight in unlicensed home child care to considerable oversight in licensed day care centres). Parents received this information prior to interacting with the conjoint survey (the control scenario was receiving no information). We investigate the influence of information on parent decision-making, including on lower-income and lower-information parents.