My research program tackles critical questions at the intersection of economics, technology, and human behavior. Through a multi-method approach that combines theoretical modeling, laboratory and field experiments, large-scale data analysis, and cutting-edge technologies like virtual reality (VR) and generative AI (GenAI), my work makes contributions to five interconnected themes that advance scholarly understanding and deliver actionable solutions to pressing societal problems.
Non-Bayesian Persuasion (with Geoffroy de Clippel), Journal of Political Economy, 130.10 (2022): 2594-2642. Online Appendix.
Abstract: Following Kamenica and Gentzkow (2011), the paper studies persuasion as an information design problem. We investigate how mistakes in probabilistic inference impact optimal persuasion. The 'concavification method' is shown to extend naturally to a large class of belief updating rules, which we identify and characterize. This class comprises many non-Bayesian models discussed in the literature. We apply this new technique to gain insight into the revelation principle, the ranking of updating rules, when persuasion is beneficial to Sender, and when it is detrimental to Receiver. Our key result extends to also shed light on the question of robust persuasion.
Online Platform and Consumer Privacy: A Bayesian Persuasion Perspective (with Jinghao Jia and Yangjie Gao).
The Impact of Personalized Product Display on Consumer Purchasing Decisions - An Experimental Study (with Jinghao Jia and Chenyang Li).
Combatting Fake Reviews with Return Insurance on E-commerce Platforms: A Signaling Game Perspective (with Ke Wang).
Cheap talk with ambiguous intention: An Analysis of Stock Recommendations.
CEO’s Accent Bias in Corporate Communication (with Da He).
Trust and Cooperation at a Confluence of Worlds: An Experiment in Xinjiang, China (with Zhe Zhang and Louis Putterman), Journal of Economic Behavior & Organization 161 (2019): 128-144.
Abstract: We study trust and willingness to cooperate among and between Uyghur and Han college students in Xinjiang, China. In our incentivized laboratory-style decision-making experiment, within and between group interactions occur among identifiable participants without traceability of individual decisions and without explicitly referencing ethnicity. We find that members of each ethnicity show favoritism towards those of their own ethnicity in both trust and cooperation and that communication enhances inter-ethnic cooperation significantly. Other results include that Uyghur and Han subjects behave differently in their willingness to cooperate relative to willingness to trust, consistent with a conjecture about cultural persistence and centuries-old differences in lifeway.
[Title omitted for peer-review] (with Fan Rao and Xiaolin Xie). Under review at the Journal of Behavioral and Experimental Finance.
Attention and Green Delivery (with Te Bao, Yeow Hwee Chua, Xuan Luo, Ruge Zhang), accepted by 2025 SGFIN Annual Research Conference on Sustainability.
The Interaction of Psychological Distance and Framing in Climate Communication: Economic Appeals versus Ecological Grief (with Rongjin Tian and Luoye Chen).
Exploring the Impact of ESG Information and Ambiguity on Asset Pricing: Laboratory Experimental Evidence (with Fan Rao and Xiaolin Xie).
Presented at the 15th Annual Meeting of the Society for Experimental Finance, Maastricht, Netherlands; and the 15th Annual Conference of the French Association of Experimental Economics, Nancy, France.
The Economic Value of Trust: Trust as a Substitute for Enforceable Contracts.
Assessing the Effects of Quantitative Green Behavior Label on Consumer Sustainable Choice (with Yannan Zhou, Luoye Chen, and Beijinni Li).
Rewiring Opportunity: How Improved Internet Infrastructure Reduces Intra-City Income Inequality in China (with Yanlin Wan, Yuzhi Hao, Aoqing Lyu, and Masaru Yarime). Under review at China Economic Review.
Abstract: This study investigates the impact of enhanced internet infrastructure on intra-city income inequality in China under the Broadband China Strategic Program (2014–2016). Employing a staggered difference-in-difference methodology, we find that the program substantially reduced intra-city income inequality of the demonstration cities, measured by a notable 0.029-point reduction in their Gini coefficients. This effect was primarily driven by income gains among low- and middle-income households, particularly in labor income. Our mechanism analysis further reveals that improved internet infrastructure facilitated new employment opportunities in the service sector and empowered low-skilled workers to transition to higher-paying roles within the platform economy. Our findings underscore the internet's potential to reduce inequality and promote inclusive labor markets, offering valuable policy insights for infrastructure investment and social equity in developing countries.
Honorable Mention, Asian Development Bank (ADB) – International Economic Association (IEA) Innovative Policy Research Award.
Previous title: Can Broadband Internet Infrastructure Narrow the Income Inequality? An Analysis of the Effects of Broadband China Strategic Program.
Generative AI as a Limited Equalizer: Experimental Evidence on Gender Gaps in Competition (with Lisha Qin and Fan Rao). Online Appendix. Under review at Experimental Economics.
Abstract: We study whether access to generative AI (GenAI) changes gender gaps in willingness to compete. In a laboratory experiment, 360 university students solve numerical and logic questions that are highly suitable for ChatGPT assistance in a Niederle and Vesterlund (2007) piece-rate versus tournament paradigm. In the main experiment (n=240), sessions are randomly assigned either to a baseline condition without ChatGPT or to a condition with ChatGPT available for optional use. ChatGPT access substantially increases performance for both women and men and largely eliminates baseline gender differences in scores, yet it does not significantly reduce the gender gap in tournament entry. Additional analyses reveal a dual-channel mechanism where ChatGPT access raises performance but lowers confidence. The gender gap in tournament entry disappears among participants with a stronger intention to use GenAI. In a follow-up information provision treatment (n=120), participants receive credible information that men and women perform similarly in prior ChatGPT-assisted tournaments. The gender gap in tournament entry persists, suggesting that correcting social-comparative beliefs alone is insufficient when uncertainty about tool reliability remains salient.
Presented at the 2025 ESA World Meeting, Beijing, China.
Platform Pricing and Exclusion Strategies: Balancing Quality, Profitability, and Social Welfare (with Yangjie Gao and Xiaotong Sun).
Presented at the INFORMS International Meeting 2025, Singapore.
When Persuasion Is Enough: Experimental Evidence on Pre-Adoption Interventions for Protective Technology (with Jie Gao and Kewei Xu).
Does AI Deepen the Divide? A Dual-Bias Perspective on AI-Generated Content and Opinion Polarization (with Aoqing Lyu).
Assessing the Value of AI Assistants in Hospitals: Efficiency Gains, Patient Experience, and Physician Perceptions (with Jie Gao and Moyan Li).
Instrumentalizing Industrial Policy for GDP Manipulation (with Zihao Qu and Guang Zhang). Under review at the Journal of Law and Economics.
Greening Like China: How Dual-Randomized Environmental Inspections Reshape Firm-Regulator Interactions and Compliance Behavior (with Da He and Yi Zhang).
From Both Sides, But Not for All: Heterogeneous Effects of a Nationwide Mental Health Policy in China (with Yanlin Wan and Yuzhi Hao). Under review at Health Economics.
Abstract: We evaluate a nationwide policy in China combining mental healthcare supply expansion with stigma reduction. Using a difference-in-differences design, we find significant average improvements in mental health. However, benefits are concentrated among low- and medium-risk populations, with no detectable gains for the high-risk. We attribute this heterogeneity to an accessibility-privacy trade-off: community-based care expansion reduces travel time but increases social visibility that disproportionately deters high-risk individuals. A discrete choice experiment confirms that high-risk individuals respond weakly to accessibility improvements due to privacy concerns. We conclude that accessibility-focused interventions can inadvertently exclude the vulnerable, necessitating privacy-protected delivery channels for high-risk groups.
Benefit Salience and Cognitive Barriers to Program Take-up: A Field Experiment (with Yanlin Wan). Under review at Behavioural Public Policy.
Abstract: Low take-up of welfare-improving programmes is often attributed to liquidity constraints, yet participation may lag even when monetary costs are negligible. We isolate the role of cognitive barriers and trust frictions in the context of China’s Rural Toilet Revolution, a unique setting where full government funding effectively rules out the inability to pay channel. To disentangle the content of the signal from its source, we implemented a 3X2 factorial RCT that cross-randomises three information frames (benefits, harms, or technical details) against the presence of a peer messenger. We document a sharp divergence between stated intentions and revealed preferences. While all interventions uniformly increased short-run self-reported willingness to participate, actual administrative enrolment responded significantly only to benefit-salience messages. We detect no effect of peer messengers. Heterogeneity analyses show that enrolment effects are concentrated among villagers with higher baseline programme literacy, suggesting a complementarity between information and prior knowledge. These results imply that even when financial constraints are removed, the inability to quantify invisible returns remains a binding constraint, and effective policy must resolve this valuation gap to trigger adoption.
Previous title: Informational Nudge for Sanitation: Experimental Evidence from China.
How Meritocratic Selection Fuels Corruption: A Laboratory Experiment on the Entitlement Effect (with Dexuan Zhang and Fan Hu).
Presented at the 2024 Asia-Pacific ESA Meeting, Singapore.
Previous title: Entitlement and Corruption: Evidence from a Laboratory Experiment.
What Determines Travelers’ Acceptance of MaaS? Insights from a Meta-Analysis (with Manlian Pan, Yipeng Chen, and Xiaotong Sun). Under review at Transportation Research Part A: Policy and Practice.
Anti-corruption Political Turnover and Government Public Service Quality (with Zihao Qu).
GenAI in School: How Guidelines and Restrictions Shape Student Performance (with Jinghao Jia and Yanlin Wan).
Abstract: We investigate the governance of Generative AI in education. Using a randomized controlled trial in two urban middle schools in China, we evaluate two policy levers: instructional AI Guidelines and Restrictions on the AI rate in writing (defined as a 30% cap on AI-generated content). We find that neither component is sufficient in isolation: guidelines without restrictions induce cognitive offloading, indicated by significantly higher restriction violation rates, particularly among male students, while restrictions alone fail to curb non-compliance. Instead, we document a strong synergistic complementarity between the two. The combination of guidelines and restrictions ensures high compliance and drives significant gains in writing performance. We conclude that a preferable policy bundle, coupling instructional support with commitment devices, effectively curbs cognitive offloading among the discipline-constrained while fostering adaptation among compliance-oriented students.
Gender Quota Policies and Workplace Sabotage (with Zhichen Guo, Yun Hou, and Dexuan Zhang).
ESG Regulation and Firms’ Voluntary ESG Information Disclosure (with Da He).
Understanding and Codesigning Photo-based Reminiscence with Older Adults (with Zhongyue Zhang, Lina Xu, Xingkai Wang, and Mingming Fan), Proceedings of the ACM on Human-Computer Interaction 9, no. 2 (2025): 1-30.
Abstract: Reminiscence, the act of revisiting past memories, is crucial for self-reflection and social interaction, significantly enhancing psychological well-being, life satisfaction, and self-identity among older adults. In HCI and CSCW, there is growing interest in leveraging technology to support reminiscence for older adults. However, understanding how older adults actively use technologies for realistic and practical reminiscence in their daily lives remains limited. This paper addresses this gap by providing an in-depth, empirical understanding of technology-mediated, photo-based reminiscence among older adults. Through a two-part study involving 20 older adults, we conducted semi-structured interviews and co-design sessions to explore their use and vision of digital technologies for photo-based reminiscence activities. Based on these insights, we propose design implications to make future reminiscence technologies more accessible and empowering for older adults.
Interpersonal Trust and Trustworthiness in a High-Immersion Virtual Environment: A Trust Game Experiment (with Hao Ling and Xiangyu Xu). Under review at the Journal of Economic Behavior & Organization.
Abstract: Understanding interpersonal trust within high-immersion virtual environments (HIVE) that underpin emerging digital platforms like the metaverse is crucial. This study conducts the first replication of the canonical trust game in a HIVE, where two participants in separate real-world spaces play an anonymous one-shot trust game in a virtual lab. Our design includes three conditions: a VR baseline group without any avatar interaction, an avatar-presence group where participants were briefly exposed to the counterpart‘s avatar visually, and a physical control group that made decisions in a traditional lab setting. We find no significant difference in trust or trustworthiness between the physical and VR baseline groups, while visual avatar presence significantly increases trust. Furthermore, male participants show higher trust propensity than females, and participants with more gaming experience exhibit greater trustworthiness. These results validate the HIVE as a powerful platform for experimental economics and offer practical insights for designing virtual environments where trust is essential.
Presented at the 2025 ESA World Meeting, Beijing, China.
Previous title: A Replication of the Trust Game in a High-Immersion Virtual Reality Environment.
Trust Formation in AI Delegation: The Interplay of Explainability and Anthropomorphism (with Chenyang Li, Zhixuan Deng, and Hao Ling). Conditionally accepted with minor revisions for the CHI Conference on Human Factors in Computing Systems (CHI’26).
Abstract: As AI agents act on behalf of users, designers increasingly combine explainability (XAI) and anthropomorphism to build trust. Yet, whether these cues create synergy or interference remains a critical, open question. Our online experiment (N=900) revealed a counterintuitive interference effect: anthropomorphism reduced trust in an explainable agent. A preregistered lab study with eye-tracking (N=57) reversed this finding: under controlled conditions, the combined design elicited the highest trust. Eye-tracking reveals the mechanism: XAI promotes deeper cognitive engagement (e.g., longer fixations), which primes users to allocate attention to social cues (e.g., avatars). Our findings show that trust depends on cognitive engagement moderating social cue processing, yielding a critical design insight: effectively pairing explanatory and anthropomorphic interfaces requires first securing the user's cognitive engagement to avoid undermining trust.
From Memory to Meaning: A Systematic Review of Reminiscence Technologies (with Zhongyue Zhang, Mingqing XU, Mengyang Wang, Lina Xu, and Mingming Fan). Conditionally accepted with minor revisions for the CHI Conference on Human Factors in Computing Systems (CHI’26).
Abstract: Technologies designed to support reminiscence, defined as the practice of engaging with one’s personal past, have become a significant area of inquiry within HCI. Although this has generated a diverse range of creative systems, the field still lacks a systematic account of the design principles that guide them. In this paper, we review 60 studies to examine both the psychosocial functions these technologies target and the mechanisms through which they operate. Our analysis suggests a predominant emphasis on positive identity construction and social connection, with comparatively less focus on functions related to everyday problem solving. To synthesize the mechanisms identified, we propose a cue-centered framework that treats mnemonic cues (e.g., photographs) as the basic unit of design. The framework organizes design mechanisms into a four-stage lifecycle: cue generation, augmentation, interaction, and sharing. It provides a conceptual vocabulary for analyzing reminiscence technologies and highlights underexplored opportunities for future research and design.
Cognitive Load as a Hidden Cost of Empathy: Trade-offs in Using VR for Inclusive Design (with Dawei Xiong, Zhijun Ma, Hao Ling, Xun Wu, and Mingming Fan). Under review at the International Journal of Human–Computer Interaction.
Abstract: Virtual Reality (VR) is used to bridge the "empathy gap" in inclusive design, yet it remains unclear whether emotionally evocative simulations improve design outcomes. To investigate this, we conducted an experiment (N = 90) comparing passive video with nonemotional and emotional VR simulations. Grounded in interviews with low-vision individuals, our interactive simulation required participants to navigate daily challenges from an embodied, first-person, low-vision perspective. We measured empathy changes with a validated scale and evaluated the quality of design ideas by their alignment with user-expressed needs, specifically auditory solutions. Our findings reveal a paradox: while VR boosted empathy, emotional stimuli may hinder design ideation, with the non-emotional group producing 145% more user-appropriate solutions. We propose a cognitive load framework specifying that affective stimuli in empathy VR may impose extraneous cognitive load, which could explain the observable trade-off between emotional immersion and analytical design capacity. This mechanistic account bridges VR empathy and cognitive load literature, offering actionable guidelines for balancing empathy with the cognitive demands of inclusive design.
Mind vs. Machine: How Cognitive Reflection and Linguistic Cues Shape Detection of AI-Generated Reviews (with Ke Wang).
Abstract: AI-generated text is increasingly embedded in everyday communication, which often appears indistinguishable from human writing. This raises urgent questions about how people discern whether text is authored by humans or large language models (LLMs). We study this challenge in the context of online product reviews, a domain where credibility directly shapes trust and decision-making. In an online experiment (N = 168), participants were recruited to distinguish 240 reviews generated by diverse LLM prompts alongside authentic human reviews. We found that although people performed above chance, they were consistently overconfident. Reviews produced through paraphrasing were especially deceptive, exploiting linguistic cues such as analytic tone, readability, social references, etc. Importantly, participants with high cognitive reflection scores were less susceptible to AI reviews across all prompt conditions. They leveraged diagnostic cues (e.g., achievement-oriented words) more effectively than intuitive participants but still exhibited blind spots. These findings reveal fundamental limits in human discernment of AI text and suggest design strategies that strengthen resilience against AI-driven misinformation.
Exploring AI Support for Life Story Interviews in Memoir Creation (with Zhongyue Zhang, Chao LIU, Luyao Shen, Wen Ku, Yuru Huang, Mengyang Wang, and Mingming Fan). Under review at the Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies.
VRization: A Framework for Designing and Implementing Virtual Reality Experiments in Experimental Economics (with Hao Ling, Zhongyue Zhang, and Mingming Fan).
Multiplayer Virtual Reality Games in Economics Education: Testing the Dual-Pathway Mechanism (Hao Ling, Chunming Ma, Xu Zhang, Lei Chen, Yi-Lung Kuo), accepted by the 2026 CAERDA International Conference.
Group Cooperation and Competition in a Highly Immersive Virtual Reality Farming Game (with Hao Ling, Chunming Ma, and Lei Chen).
The Gendered Pitch: Causal Evidence of Bias in Investor Judgments from an AI Face-and-Voice Manipulation Experiment (with Yun Hou, Hao Ling, and Xiangyu Xu).
Time Perception and Intertemporal Choice in VR (with Aoqing Lyu, Hao Ling, and Xiangyu Xu).
The Effect of Tangible Rewards and Automated Smart Contracts on Present Bias (with Chongye Huang and Chenyang Li).
Abstract: We examine how retail investors’ topic-specific attention evolves and predicts stock returns and volatility. We construct firm-day attention measures using topic modeling on 340 million posts from China’s largest stock forum. We identify four recurring topic lifecycles and increased discussion regularity after 2015. Attention to “Continuous Stock Drop” predicts lower next-month returns with partial reversal in one to two months, while attention to “Capital Market Reform” forecasts higher volatility persisting up to nine months. These results are stronger among smaller firms, non-main-board firms, and in the post-2015 period, and are robust to orthogonalizing against recent returns and alternative specifications.
Beyond Aversion and Seeking: Examining Divergent Ambiguity Attitudes in Experimental Asset Markets (with Fan Rao).
Presented at the 2024 Theoretical and Experimental Macroeconomics Workshop, Singapore.
The Influence of Pay Frequencies on Consumption and Saving (with Lisha Qin).
Is There a Founder Penalty in Financial Job Markets? An Audit Study in China (with Kingking Li and Chi Zhang).
Categorical thinking and life experiences: A metatheory of anomalies in economic lab experiments (with Chongye Huang and Chenyang Li).