Student Researchers are a vital part of the Social Science research community.

Class 2016

Angelo Kamus

When social sanctions turn against society: An experiment on the long-run costs of social sanctions

Abstract:Social sanctions are often thought to support welfare-enhancing norms through cooperation, but certain social sanctions may inhibit adaptation to changing environments. They may enforce negative social norms, such as the case of discrimination against uncircumcised girls in parts of Africa and the Middle East. Individual preferences change over time, and when a sufficient number of people change their preferences, it should be an optimal break out point at which social change is predicted to occur. Social sanctions, however, may delay that process. Can social sanctions lock groups in conformity traps? What are some of the conditions which increase the likelihood of this happening? In our experiment, we pair subjects together in a matching group, wherein each subject selects one of two choices, Blue or Green. Players begin as Type A subjects, with a certain rate of change that they will switch to Type B. The different treatments manipulate whether their rate of change and/or their type is known. Their payoffs are determined by their types, their color choice, and the color choice of their matched participants. In addition, participants have an option to reduce the earnings of their matched participant depending on the color they choose. We want to observe if social sanctions, through an earnings reduction, delay the period at which social change is predicted to occur.

Lucia Salem

Ranks or scores: does giving ordinal instead of cardinal rank feedback deter sabotage?

This paper studies unethical behavior under a flat-wage scheme in a lab experiment. I seek to investigate the role of status seeking on effort and sabotage under two different settings: giving cardinal ranks versus giving ordinal ranks. The experiment is meant to mimic a real company work setting and will determine which of these two incentive mechanisms induces the most effort and the least sabotage. I expect to find less sabotage and more effort when relative performance feedback is given in the form of ordinal ranks rather than cardinal scores.

Class 2015


Henry Bradley

Communication and Coordination in Games with Third-Party Externalities

    In many cases the actions of decision makers in coordination problems do not solely impact personal welfare, but also the welfare of third parties. Standard models of agent’s preferences, which do not consider the welfare of others in utility functions, predict agents to behave in a rationally self-maximising way. That is, even in the light of potentially large negative externalities, the decision maker will behave according only to his own payoffs, without any weight given to the outcomes for other parties. Evidence from Bland and Nikiforakis (2013) demonstrates that participants in an experiment are willing to act in a way that increases their income slightly, even if doing so causes substantial negative externalities and reductions in total group welfare. Elsewhere, studies have shown that communication can improve group welfare outcomes, however, in these studies, increases in group welfare did not require a sacrifice in the earnings of decision-makers. Hence, it remains unclear whether communication will have the same effect when increasing total surplus implies a reduction in earnings of the decision-makers. Given this evidence for the positive impact of the introduction of communication opportunities on group welfare, this capstone project sheds light on whether communication can move behaviour away from the tendency to anti-social consequences in coordination problems as shown to be acute in Bland and Nikiforakis (2013). If communication opportunities do induce pro-social behaviour, the project will seek to answer why this is so. To that end, using an experimental set-up I investigate the impact of communication opportunities on the outcomes of a strategic coordination game featuring a pro-social and anti-social equilibrium.


Marek Mihok

Avenge or let be? Study of punishment retaliation by third (i.e. unaffected) and second parties.

    The goal of the project is to explore whether people react differently to punishment by second or by third parties. In particular, I will investigate whether subjects are more willing to counter-punish third parties than second parties.Punishment can be meted out by second parties (i.e., individuals are directly affected by the violation of the norm) or third parties (i.e., individuals are not directly affected by the violation of the norm). Previous studies have shown that even if punishment is costly to unaffected third parties, the latter are still willing to incur a cost to punish selfish behavior.This Capstone Project will provide the first empirical evidence comparing the extent to which second and third parties are counter-punished. This is important as, if individuals are less willing to retaliate third-party punishment, the evidence could provide a rationale for the emergence of central (impartial) authorities in charge of enforcing social norms and cooperation.


Anirudh Sood

Raising Funds 101? Groups, Individuals, and the "Identified Victim" Effect

    Many NGOs rely on empathic arousal to raise funds. A common source of such arousal is a victim’s story and image. In my paper, I study individual and group contributions made towards helping Moke, a 7-year-old child living in Democratic Republic of Congo, suffering from a dermatological condition. In my experiment, I control levels of identification by offering the image, name, and age of the victim to the treatment. My experiment confirms the existence of the “Identified Victim” effect in individuals: individuals contribute more towards helping an identified victim than an unidentified victim. Importantly, I extend the study to groups to find, surprisingly, that groups contribute more towards helping an unidentified victim than identified victim. Finally, the results also demonstrate that groups are less rational than individuals i.e. groups contribute more towards helping the unidentified victim than individuals.

Yuqi Sun

How Do Information Effects Affect Entrepreneurial Behaviours? Empirical Evidence

    This research aims to understand how information effects affect individuals' decision making, individuals' productivity, and market outcome in a market setting. Information effects are defined as the knowledge of the highest-profit-earning participants’ profit-making performances and decisions. Market outcomes are defined as the market equilibrium, market prices, and individual's entrepreneurial decisions. Mas & Moretti (2009) and Falk & Ichino (2006) show that peer effects improve productivity in particular settings, especially when subjects observe each other, but there are questions to be asked about the effects in a market setting. To investigate the subject, I conducted lab experiments that simulates a capital market, in which participants can rent or lend capital at a market price, and make profits by performing real-effort tasks with their remaining capital. I expect to show that making the knowledge of best profit-making practices public encourages individuals' strategic decision making, and improves individuals' productivity.

Class 2014


Sidak Gebre Yntiso

Self-Confidence in Information Processing: Experimental Methods and Evidence

Do individuals interpret ambiguous feedback optimistically or pessimistically when their self-esteem is at stake? The standard finding from numerous social psychology and behavioral economics studies is overconfidence. This paper demonstrates an experimental situation which garners the opposite effect; here, people over weigh negative performance feedback which leads to overly-pessimistic beliefs about ability. In addition, I introduce a number of methodological improvements (in performance test, groups and signal structure). This paper also compares belief accuracy across two types of belief elicitation mechanisms - the quadratic scoring rule and the crossover mechanism. There is no significant difference in belief accuracy between the two mechanisms.