Research Overview

My research focuses on two areas of the digital economy. The first aims to understand how people interact with different types of information, including personal, instrumental, and misinformation. This line seeks to provide insight into the effectiveness of limited regulations that allow for the free use, sharing, and sale of information.

The second line is to discover (non-monetary) performance boosters for lone workers in various contexts, drawing on insights from behavioral sciences.

In summary, my work involves systematic analyses of information usage and highlights the potential benefits or consequences of different information policies.

Publications

Yi-Shan Lee and Roberto A. Weber (2024). "Revealed Privacy Preferences: Are Privacy Choices Rational?" Management Science, forthcoming. 

Abstract: We investigate the extent to which tradeoffs involving the sharing of personal information exhibit consistency with an underlying rational preference for privacy. In an experiment, people engage in tradeoffs across two domains of personal information, allowing us to classify whether their choices satisfy the Generalized Axiom of Revealed Preference. Sixty-three percent of subjects act consistently with a rational preference ordering when allocating privacy levels, despite substantial heterogeneity of privacy attitudes. Individuals who are inconsistent when engaging in such privacy tradeoffs exhibit substantially more costly preference reversals when pricing the sharing of their personal information. Our results imply that allocating privacy property rights by monetizing personal information can have distinct monetary welfare consequences for people with different degrees of rationality in their underlying ability to make sensible tradeoffs involving personal information sharing.  We also provide evidence that preferences elicited over choices in our experiment correlate with real-world privacy behaviors.

Wei Chen, Shu-Yu Liu and Chih-Han Chen and Yi-Shan Lee (2011). “Bounded Memory, Inertia, Sampling and Weighting Model for Market Entry Games”  Games, 2(1), 187- 199 

Abstract: This paper describes the “Bounded Memory, Inertia, Sampling and Weighting” (BI-SAW) model, which won the http://sites.google.com/site/gpredcomp/Market Entry Prediction Competition in 2010. The BI-SAW model refines the I-SAW Model (Erev et al. [1]) by adding the assumption of limited memory span. In particular, we assume when players draw a small sample to weight against the average payoff of all past experience, they can only recall 6 trials of past experience. On the other hand, we keep all other key features of the I-SAW model: (1) Reliance on a small sample of past experiences, (2) Strong inertia and recency effects, and (3) Surprise triggers change. We estimate this model using the first set of experimental results run by the competition organizers, and use it to predict results of a second set of similar experiments later ran by the organizers. We find significant improvement in out-of-sample predictability (against the I-SAW model) in terms of smaller mean normalized MSD, and such result is robust to resampling the predicted game set and reversing the role of the sets of experimental results. Our model's performance is the best among all the participants.

Working Papers 

"A Sticky Threat: How a Single Exposure of Misinformation Changes Beliefs, Behaviors, and Perceived Norms" 

Reject and Resubmit at Management Science.

with Erin Krupka 

Abstract: This paper demonstrates the social cost of misinformation by its impact on behavior and perceived norms. Specifically, we ask whether a single exposure to false information can change not only beliefs but also behavior and social norms. To answer this question, we conduct two experiments on Amazon's Mechanical Turk, examining the effects of a false negative claim about either a 2018 migrant caravan or an endangered species on donation behavior (to a charity supporting each cause respectively), beliefs, and social norms. Our results indicate that a single exposure to misinformation lowers the average donation by 32% in both experiments. The false claim reinforces incorrect prior beliefs and changes individuals' perceptions of social norms. Interestingly, providing debunking information fails to restore either behavior or norms to baseline levels. Our findings demonstrate that misinformation is a sticky threat, making it difficult to undo its harmful effects on behavior and social norms. These results have important implications for how misinformation should be regulated and how best to attenuate its impact.

“Long-shot Incentives in Dynamic Winner-take-all Contests: An Experimental Study” (link to PDF)

with Joseph Tao-yi Wang 

Abstract: Substantial empirical evidence suggests that winner-take-all incentives encourage risk-taking. We experimentally investigate how risk-taking varies across interim positions and remaining time in a 24-period winner-take-all contest mimicking real-life tournaments, in which leading players can be taking more risk when close to the top, while end-game behaviour shifts strategically. In every period, each participant independently commits to either a long-shot or piece-rate scoring policy, and then competes against an opponent according to this self-imposed policy.  The equilibrium policy decision in this long-horizon contest is non-monotonic in both dimensions and especially volatile for interim leaders. Our experimental results show that equilibrium predictions fully mediate the effect of winning rates on risk-taking. Consistent with aggregate equilibrium predictions, trailers take more long shots than leaders, and long-shot adoption is encouraged by larger long-shot gains. This result suggests that measures such as extending patent life are effective for encouraging path-breaking innovation. Nevertheless, trailers, instead of leaders, take more long shots toward the end. Deviations from equilibrium indicate that contestants may follow score-maximizing heuristics early on, but adhere to winner-take-all predictions when approaching the end-game. Gender and risk preferences play secondary roles in long-shot adoption.

Works in Progress

“Acquisition and Utilization of Information in Social Networks” 

with Ningning Cheng and Feng Qin

Abstract: We experimentally investigate information acquisition and utilization in small social networks. A group of subjects each observe a noisy signal about an unknown state, make costly connections to acquire each other's signals, and utilize available information to choose an action commensurate to the state and average action. Our findings support two predicted monotonicity properties and document a set of behavioral anomalies. As predicted, subjects acquire signals from those who are more popular, and those who are themselves more popular acquire fewer signals. Relative to equilibrium predictions, however, subjects persistently under-acquire information and suboptimally utilize available information by, notably, ignoring otherwise useful, costly acquired signals. Furthermore, the poor utilization of information intensifies as more information is acquired. We explore plausible explanations for these findings as a step toward admitting the observed anomalies into the theoretical domain.

“Personal Information and Information as Public Goods” 

with Josie Chen

Abstract: We design an experiment to compare the public-good generating process and social welfare between the information economy and manufacturing economy. The experiment has three conditions in a repeated environment: personal information (PI), information (I), and money (M). In both PI and I conditions, personal data is monetized: it can be kept as private goods or used for generating public goods with monetary values that in-group members will share. The M condition is the classic public goods game using monetary units. 

“Advice Giving—a Performance Booster for Remote Workers?” 

with Feng Qin, Outstanding Research Paper Award at China BEEF Conference (2024)

Abstract: How can remote employees working outside a traditional office environment be effectively motivated? This paper investigates the feasibility of using advice-giving as a performance booster by conducting a controlled experiment. Remote workers are randomized into one of the three conditions: baseline, advising others and self-advising, and perform a real-effort task involving trading off between cost and benefit of input effort. Our results show that self-advising increases workers’ performance by 10%; this effect size is nearly twice as much as advising others, a renowned approach for boosting performance (Eskreis-Winkler et al., 2018, 2019). The treatment effect mainly comes from the performance improvement of workers who initially scored at the bottom half. We examine potential mechanisms underlying the effects and find that a rise in self-confidence is associated with performance improvements. These findings suggest that self-advising can be a powerful tool for increasing the confidence and performance of workers that managers and policymakers can utilize in remote workplaces.