Behavior Working Group

This study investigates the spread of misinformation on social networks using an experimental approach to isolate and analyze key factors influencing rumor diffusion. We examine two settings: one where individuals can only forward or block messages and another where they can also modify the message content. Participants act as unbiased agents seeking the truth, interacting within a network of both biased and unbiased agents. The biased agents favor a specific outcome, aiming to sway network beliefs. The experiment tests various network scenarios, altering participants’ positions and their neighbors’ types to assess how message credibility and transmission behavior change based on message source proximity and sender type. Through theoretical predictions, we explore agents’ transmission decisions, focusing on their responses to messages of differing credibility across varied network distances. Our findings are expected to shed light on how misinformation spreads within networks, particularly how unbiased individuals handle messages from biased sources and whether network filtering can effectively curb rumor propagation. This study contributes to understanding the complex dynamics of misinformation diffusion, including potential biases that may inadvertently facilitate the spread of false information.

The aging of the general physician workforce in developed countries is expected to lead to increased practice closures. Hence, concerns arise regarding the health effects of such closures, particularly for patients facing them during critical life stages such as pregnancy. However, no study exists to date on the health effects of general physicians’ (GP) practice closures during pregnancy. I assess the effects of such closures during pregnancy on birth outcomes in Denmark. I compare the birth outcomes of mothers experiencing practice closures within nine months post-conception to those facing closures nine months pre-conception. I find a small to medium-sized adverse effect of discontinuity in care on birth outcomes. Closures increase the share of births of fetuses who are small for their gestational age. Such result hints at closures harming the neonatal health of marginalized children. The negative effect on birth weight is especially pronounced when the closure happens in the last trimester of pregnancy. Consistently, mothers affected by GP practice closures during pregnancy experience small disruptions in healthcare provision at the extensive and at the intensive margin.

This study aims to investigate the psychological and social factors influencing populist support in France through experimental methods. Using a factorial design, participants were exposed to controlled scenarios to assess in-group and out-group dynamics, including money-burning tasks and perceptions of the rule of law. The research explores how political alignment, trust in law, and emotional triggers shape voter behavior, particularly toward populist and non-populist ideologies. Results are expected to elucidate the interplay of social identity and perceptions of justice in voting decisions, providing insights into the rise of populism and its implications for political systems. The findings aim to contribute to the broader discourse on voter behavior, emotional decision-making, and the rule of law in contemporary democracies.

“Losers’ consent”, i.e., the acceptance of political outcomes by those for whom they are undesirable, is central to political stability and democratic viability. The present project investigates the undocumented impact of a specific institutional factor, the voting rule, on electoral consent. Existing research substantiates that alternative voting rules like Approval Voting – under which voters must decide whether they approve each candidate – encourage political expression, attenuate electorate polarization and designate more consensual winners. We propose a novel experimental design to assess whether such voting rules could have a positive impact on voters’ acceptance of democratic decisions. Using an online electoral setting where participants vote to allocate money to a French political party, we observe incentivised voting behaviour and electoral consent. We compare two randomly allocated voting rules, First-past-the-post and Approval voting, to investigate their effect on losers’ democratic compliance.

This paper investigates whether female politicians face constraints on emotional expressions and quantifies their electoral consequences. Using a comprehensive dataset of candidate manifesto images from the 2022 and 2024 French parliamentary elections, we first establish two empirical facts: (i) female candidates are 19 percentage points more likely to display positive emotions than male counterparts and (ii) female candidates not displaying positive emotions receive lower vote share, while comparable male candidates do not. Second, we implement a randomized experiment to study the causal effects of gendered emotional expressions. We generate images of politicians with varied facial expressions using artificial intelligence methods. We have designed an experimental protocol where we plan on asking subjects to evaluate identical candidates displaying different emotional expressions. The design enables us to precisely measure how the perception of women changes when they violate emotional display expectations.