Main
Opinion Formation under Endogenous Social Incentives
(with Fabian Dvorak, Urs Fischbacher, and Katrin Schmelz)
Prepared for submission
Abstract: We experimentally isolate genuine change from strategic misrepresentation of publicly expressed opinions in a setting that mimics key features of social media environments. We document that people adopt majority opinions even when their views remain private, which substantiates widespread concerns that echo bubbles foster within-group homogeneity and thereby accelerate polarisation across groups. We further show that endogenous social evaluation, a defining feature of social media platforms, causes additional conformity to predominant opinions that is not backed by genuine preference alignment. Instead, such strategic conformity is motivated by the aim to attract social reward or avoid social sanctions. The effects we find are robust across the entire set of topics we consider and large in magnitude. Importantly, the coexistence of both effects indicates a causal role of social feedback tools in the formation of echo chambers and polarisation, operating through a self-reinforcing mechanism whereby illusory consensus amplifies intrinsic conformity.
Presented at: FSB Research Seminar Innsbruck, EEA ESEM Meeting Rotterdam, ESA World Meeing Lyon, IAREP-SABE Nice, Spring Summit Innsbruck, briq Bonn, Tinbergen Institute Amsterdam, Nottingham, Konstanz
Social Closeness and Peer Effects
(with Simon Gächter and Silvia Sonderegger)
Working Paper in preparation
Abstract: We study how social closeness causally shapes peer effects in cooperative behaviour. While peer influence is typically modeled as uniform, we argue that identical information can have heterogeneous effects depending on the perceived social closeness between observers and observed peers. We test this hypothesis in a one-shot public goods experiment that allows peer-to-peer observation across payoff-independent groups. A key contribution of our study is the combination of a novel experimental design that exogenously varies social closeness between peers with structural equation modeling and the benchmarking of behaviour using ex ante elicited cooperation preferences (ABC method). This approach allows us to disentangle preference-based and belief-based channels in peer effects. We find that social closeness fundamentally alters the mechanisms of peer influence. Behavioural contagion from observing peers is strongest under high social closeness, consistent with enhanced social learning. In contrast, reputational effects of being observed are pronounced under anonymity but vanish when social closeness is high. Belief-based channels affect cooperation but are not strongly moderated by social closeness.
Presented at: ESA Osaka, FAEE Dijon, TIBER Tilburg, RExCon Moscow, IAREP-SABE, Nottingham, SABE, Konstanz, Bogotá EEC, Vienna, NoBeC U Pennsylvania (Poster), ESA Dijon, Essex CDEC, Spring Summit Innsbruck, MPI Berlin
Online Behaviour and Opinion Polarisation
Data collection ongoing
Presented at: Institute for Advanced Studies Vienna, Nottingham
Misperceived Polarisation and Large-Scale Cooperation
(with Simon Gächter)
Data collection ongoing
Subjective Views, Objective Truths, and the Emergence of Pluralistic Ignorace
Pilot Data collected
Signaling, Collective Misperception, and Stigma in Mental Health Service Take-up
Design phase
Social Media and Mental Health: Self-Selection and Naiveté
Design phase
Methodology and Replicability
Comparing Human-Only, AI-Assisted, and AI-Led Teams on Assessing Research Repro ducibility in Quantitative Social Science
(with Brodeur A., Valenta D., Marcoci A. et al)
R&R PNAS
Computational Reproducibility and Robustness of Empricial Economics and Political Science Research Between 2022 and 2023
(with Brodeur A., Mikola D., Cook N. et al)
R&R Nature
Assessing Robustness to Varying Clustering Methods and Samples in Ambuehl, Bernheim, and Lusardi (2022): Replication and Sensitivity Analysis
(with Chi Danh Dao, Guidon Fenig, and Jin Young Yoon)
I4R Discussion Paper Series No. 110
Predicting Reproducibility and Robustness
(with Institute For Replication)
Working Paper in preparation