PEER-REVIEWED PUBLICATIONS

Kane, J.,  Jangili, S., Sandor, J., and Sgorlon, V. “Evolution by National Election? Partisanship, Voting Behavior, and Sentiments toward Marginalized Groups” 

Forthcoming in Politics, Groups, and Identities (alphabetically-ordered authorship).

WORKING PAPERS

Sgorlon, V. “Accountability and Political Survival after Brazil’s Lava Jato” 

Under review. 


Abstract: While democratic theory assumes citizens sanction malfeasance, empirical evidence remains mixed. I combine a quasi-natural experiment with survey analysis to exploit Brazil’s Operation Lava Jato to test the link between corruption and voter behavior. First, analysis of survey data fielded immediately after a scandal identifies the behavioral mechanism. I show that reported levels of information about the scandal drive voters to punish exposed candidates in hypothetical election scenarios. Second, using a Doubly Robust Difference-in-Differences estimator, I estimate the effect of public exposure of candidates on real-life electoral outcomes, and find that exposed candidates survive electorally. These findings highlight the differences in the impact of corruption between hypothetical and real electoral scenarios.


FINAL-STAGE WORK IN PROGRESS (full drafts available)

Urbina, I. and Sgorlon, V. “Fractured Trust: The Spillover Effects of Police Violence on Political Trust and Democracy”.


Abstract: How does police violence impact citizens’ trust in democratic institutions? This paper explores the influence of police interactions, particularly through exposure to violence, on citizens’ satisfaction with democracy and trust in its institutions. We conducted three studies: an experimental study in Brazil, a causal inference design in Chile, and an observational analysis using LAPOP data across various Latin American countries. Our results reveal that, overall, experiences of police abuse erode satisfaction with democracy and trust in key democratic institutions. Furthermore, race, political identity, and trait aggression often moderate the effects of exposure. Our findings highlight the critical impact of police violence on democratic support and suggest that addressing such violence is crucial for sustaining democratic legitimacy, particularly in regions experiencing democratic backsliding. 


Sgorlon, V. “The Importance of Left-Right Ideology as an Identity in Latin America”.*


Abstract: Models of partisanship as a durable social identity travel poorly to contexts where attachments are weak or absent for much of the electorate due to unstable party systems. I argue that in such contexts, ideological identity—a psychological attachment to broad ideological camps—can serve as the primary predictor of political behavior. Using panel datasets, Study 1 shows that left-right identification persists over time as well as or better than partisanship in Argentina, Brazil, and Chile, but not in Mexico, which has a more stable party system. Study 2 finds that ideological group attachment (AIG) predicts vote congruence, political interest, and political engagement independently of other political identities. Study 3, a pre-registered experiment in Brazil, Chile, and Mexico, randomized party ideological labels and policy: high-AIG respondents in Brazil and Chile continue to favor ingroup-labeled party profiles even when doing so requires accepting policy-incongruent platforms, and show a blunted—though not eliminated—responsiveness to policy congruence relative to low-AIG respondents. AIG thus functions as an identity-based bias that systematically pulls party evaluation toward ideological ingroups, operating independently of partisan attachment and in parallel with, rather than strictly overriding, substantive policy considerations. Together, the studies establish ideological group attachment as a more stable and behaviorally consequential identity than partisanship across countries in the region with unstable party systems.


Sgorlon, V. and Kline, R. "Who Loves and Who Hates? Disentangling Affective Polarization in Brazil".*


Abstract: Affective polarization is usually attributed to partisan identity, yet in Brazil partisanship explains only part of the country's political division. We compare three politicized identities at once, left-right ideology, identification with the Workers' Party and its rejection (petismo and antipetismo), and attachment to Lula and Bolsonaro, using a survey of 1,228 Brazilians and a pre-registered triadic give-or-take dictator game with real stakes. Our design tests three affective polarization behaviors: ingroup favoritism, outgroup deprivation, and active harm toward the outgroup. Favoritism is present and similar across all identities, as identification alone should predict under Social Identity Theory. Taking money from the outgroup roughly doubles as the object of identification narrows from ideology to leader attachment or party, and the escalation comes from more people choosing to take rather than from takers taking more. What a citizen's identity is attached to determines how far they will go against the other. In weak-party systems, measuring polarization through partisanship alone can understate the conflict and miss where hostility actually attaches.


Senturk, A. E., Sgorlon, V., Echeverri, A., and Urbina, I. “The Economy or My Economy? Unpacking Immigration Threat in the United States” (shared first-authorship).

Abstract: Citizens frequently misperceive the size of immigrant populations, but whether correcting these misperceptions changes attitudes remains unclear. We examine whether the consequences of corrective information depend on the context it refers to and on the direction of respondents' prior misperceptions. In a preregistered survey experiment with 1,641 U.S. respondents, participants received accurate information about the number of immigrants in the workforce nationally and in their state, in their own occupation and occupational group, both, or neither. Respondents updated their beliefs toward the true values, but effects on perceived economic threat or immigration policy preferences are heterogeneous. Underestimators reported heightened threat after correction, while large overestimators became less worried about losing their jobs and, when corrected about national and state figures, more supportive of legalization pathways. Muted effects of corrective information may thus partly reflect opposing responses among respondents misperceiving in different directions, and a correction that is not personally relevant.


* Part of the dissertation