About

Ágnes Horvát is an Associate Professor whose research aims to make the hyperconnected and AI-infused Web more efficient, equitable, and inspiring for scientists, entrepreneurs, and creatives. Capitalizing on her background in physics, computer science, film, and media, her work employs a multidisciplinary approach that includes large-scale data analyses, online experiments, behavioral surveys, and AI to uncover how online spaces operate and how they can be designed to better serve people.

Research topics

• Computer Science
• Political Science
• Psychology
• Internet privacy
• Medicine
• Artificial Intelligence
• Public relations
• Philosophy
• World Wide Web
• Data science
• Epistemology
• Psychoanalysis
• Social psychology
• Communication
• Cognitive psychology
• Law
• History

Selected publications

• “They Only Silence the Truth”: COVID-19 Retractions and the Politicization of Science

  2024

  Senior authorCorresponding
  • Political Science
  • Computer Science
  • Internet privacy

  Retracted COVID-19 articles have circulated widely on social media. Although retractions are intended to correct the scientific record, when trust in science is low, they may instead be interpreted as evidence of censorship or simply ignored. We performed a content analysis of tweets about the two most widely shared retracted COVID-19 articles, Mehra20 and Rose21, before and after their retractions. When Mehra20 was seen as a politicized attack on Donald Trump and hydroxychloroquine, its retraction was broadly shared as proof that the article had been published for political reasons. However, when Rose21 was seen as evidence of vaccine harm by vaccine opponents, its retraction was either ignored or else framed as a conspiracy to censor the truth. These results demonstrate how retractions can be selectively used by scientific counterpublics to reframe the regulation of science as evidence of its institutional corruption.

  DOI

• Characterizing Online Media on COVID-19 during the Early Months of the Pandemic

  Journal of Quantitative Description Digital Media · 2021 · 2 citations

  Senior authorCorresponding
  • Computer Science
  • Political Science
  • Computer Science

  The 2019 coronavirus disease had wide-ranging effects on public health throughout the world. Vital in managing its spread was effective communication about public health guidelines such as social distancing and sheltering in place. Our study provides a descriptive analysis of online information sharing about coronavirus-related topics in 5.2 million English-language news articles, blog posts, and discussion forum entries shared in 197 countries during the early months of the pandemic. We illustrate potential approaches to analyze the data while emphasizing how often-overlooked dimensions of the online media environment play a crucial role in the observed information-sharing patterns. In particular, we show how the following three dimensions matter: (1) online media posts’ geographic location in relation to local exposure to the virus; (2) the platforms and types of media chosen for discussing various topics; and (3) temporal variations in information-sharing patterns. Our descriptive analyses of the multimedia data suggest that studies that overlook these crucial aspects of online media may arrive at misleading conclusions about the observed information-sharing patterns. This could impact the success of potential communication strategies devised based on data from online media. Our work has broad implications for the study and design of computational approaches for characterizing large-scale information dissemination during pandemics and beyond.

  DOI

• Network Structures of Collective Intelligence: The Contingent Benefits of Group Discussion

  arXiv (Cornell University) · 2020 · 1 citations

  Senior authorCorresponding
  • Computer Science
  • Artificial Intelligence
  • Social psychology

  Research on belief formation has produced contradictory findings on whether and when communication between group members will improve the accuracy of numeric estimates such as economic forecasts, medical diagnoses, and job candidate assessments. While some evidence suggests that carefully mediated processes such as the "Delphi method" produce more accurate beliefs than unstructured discussion, others argue that unstructured discussion outperforms mediated processes. Still others argue that independent individuals produce the most accurate beliefs. This paper shows how network theories of belief formation can resolve these inconsistencies, even when groups lack apparent structure as in informal conversation. Emergent network structures of influence interact with the pre-discussion belief distribution to moderate the effect of communication on belief formation. As a result, communication sometimes increases and sometimes decreases the accuracy of the average belief in a group. The effects differ for mediated processes and unstructured communication, such that the relative benefit of each communication format depends on both group dynamics as well as the statistical properties of pre-interaction beliefs. These results resolve contradictions in previous research and offer practical recommendations for teams and organizations.

Frequent coauthors

• Thanasis Vergoulis

  Athena Research and Innovation Center In Information Communication & Knowledge Technologies

  1 shared

• Abdullah Almaatouq

  Massachusetts Institute of Technology

  1 shared

• Angelo A. Salatino

  The Open University

  1 shared

• Feng Xia

  RMIT University

  1 shared

• Daniel M. Romero

  1 shared

• Andrea Mannocci

  1 shared

• Misha Teplitskiy

  University of Michigan–Ann Arbor

  1 shared

• Ying Ding

  Shanghai University of Electric Power

  1 shared

Labs

• Lab on Innovation in Networks and KnowledgePI

Awards & honors

• NSF CAREER
• CRII
• Northwestern Presidential Fellowship

Similar researchers at Northwestern University

• Mercouri G. Kanatzidish-205
• Chad Mirkinh-189
• Omar Farhah-184
• Tobin J. Marksh-176
• Joseph Hupph-171