Research
My research interests are strategic communication, social information, and social norms.
Publications
Chang, D., E. L. Krupka, E. Adar, A. Acquisti, “Design, Perception, and Action: Norms Shaping Designs." In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems. 2016
Nudging behaviors through user interface design is a practice that is well-studied in HCI research. Corporations often use this knowledge to modify online interfaces to influence user information disclosure. In this paper, we experimentally test the impact of a norm-shaping design patterns on information divulging behavior. We show that (1) a set of images, biased toward more revealing figures, change subjects’ personal views of appropriate information to share; (2) that shifts in perceptions significantly increases the probability that a subject divulges personal information; and (3) that these shift also increases the probability that the subject advises others to do so. Our main contribution is empirically identifying a key mechanism by which norm-shaping designs can change beliefs and subsequent disclosure behaviors.
Chang, D., R. Chen, E. L. Krupka, “Rhetoric Matters: A Social Norms Explanation for the Anomaly of Framing.” Games and Economic Behavior. 2019
Ample evidence shows that certain words or ways of phrasing things can cause us to change our preferences. We demonstrate one mechanism for why this happens - "framing" evokes norms which then influence choice. We use a laboratory study to test the impact of describing a series of dictator games with either politically charged tax or neutrally-framed language. Subjects’ political identities interact with these frames, causing changes in both norms and choices. Framing makes Democrats prefer equalized outcomes, and Republicans reluctant to redistribute payments even when it leaves them disadvantaged.
Researchers in the social sciences have increasingly studied how emotions influence decision-making. We argue that research on emotions arising naturally in real-world environments is critical for the generalizability of insights in this domain, and therefore to the development of this field. Given this, we argue for the increased use of the “quasi-field experiment” methodology, in which participants make decisions or complete tasks after as-if-random real-world events determine their emotional state. We begin by providing the first critical review of this emerging literature, which shows that real-world events provide emotional shocks that are at least as strong as what can ethically be induced under laboratory conditions. However, we also find that most previous quasi-field experiment studies use statistical techniques that may result in biased estimates. We propose a more statistically-robust approach, and illustrate it using an experiment on negative emotion and risk-taking, in which sports fans completed risk-elicitation tasks immediately after watching a series of NFL games. Overall, we argue that when appropriate statistical methods are used, the quasi-field experiment methodology represents a powerful approach for studying the impact of emotion on decision-making.
Working papers
Chang, D. “Talk is cheap, but cheap talk speaks volumes.” Manuscript.
Communication, even cheap talk, can mitigate the impact of moral hazard when the outcome is a noisy indicator of the agent's effort. However, it is unclear whether and how such cheap talk can influence the principal's beliefs and decisions when the principal has prior information about the agent. In this paper, I test the extent to which an agent's cheap talk response to credible, but noisy information about himself or herself can influence the principal's beliefs and subsequent choice of an agent. I further test whether different types of cheap talk responses might be more effective given the principal's knowledge of agents' relative past performances. In an experiment, I vary the information that the principal receives about the relative past performance of two agents in an earlier task. The principal (the "manager") then receives the two agents' (the "candidates") costless, unverifiable responses to the information (cheap talk responses) before choosing to hire a candidate for a new task. These responses may either be self-promoting ("sweet"), other-defaming ("mean") or a neutral statement ("neutral"). I show that such cheap talk affects the principal's beliefs of an agent's past performance and the principal's choice in choosing an agent to act on his or her behalf. Moreover, I nd that the impact of different types of responses on beliefs and choice depends on the prior information that the principal possesses about the agent. Lastly, I show evidence that these cheap talk responses can influence the principal's decision beyond altering the principal's beliefs about the agent's past performance.
D. Chang, Acquisti, A., E. Adar, E. L. Krupka. "Norms- and Non-Norms Nudges for Behavior Change." Submitted.
The potential of nudges to influence behavior with minimal intrusion has led to the creation of "nudge units" in government and to recent academic and commercial experimentations. However, the effective implementation of nudges is limited by the fact that only a few studies directly compare the impact of different types of nudges and by the "file drawer" problem in meta-analysis. To test the relative effectiveness of nudges at influencing behavior, we conduct an experiment that compares the relative effectiveness of two norm-nudges ("Descriptive Norms" and "Reciprocity") against two non-norm nudges ("Personalization" and "Time Limit") in motivating reading content that contains news which is incongruous with an individual's political ideology. We further test whether a human-selected nudge (by someone who knows the individual) is more effective than a randomly allocated one. We find that the "Reciprocity" nudge is the most effective in motivating interaction with the email and that individuals correctly identify this nudge as the most effective of the four. However, human-selected nudges do not outperform nudges chosen at random in motivating interaction with the email. This research produces a direct comparison of the relative effectiveness of norm and non-norm nudges and shows that, in this context, reciprocity has a modest effect on behavior change and that human-selected nudges that rely on personalized knowledge of the target fare worse. The results suggest that nudge units should be cautiously optimistic regarding the impact of nudges to achieve behavior change.
Work in progress
Chang, D., R. Chen, E. L. Krupka. “Identifying Norms for Specific Groups.”
People from across the political spectrum struggle to understand each other’s values and moral systems (Graham et al., 2012; Haidt and Graham, 2007). Many claim that this division stems from political parties misjudging each other’s values. We test the degree to which subjects across the Democrat/Republican divide misjudge each other’s norms. We extend the norms elicitation protocol in Krupka and Weber (2013) in two ways. In Krupka and Weber, subjects play a pure matching coordination game in which their goal is to anticipate the extent to which a randomly drawn match will rate an action as socially appropriate or inappropriate. We modify this to explicitly identify the political affiliation (Democrat or Republican) of the reference group and ask subjects to guess the modal rating of the subjects who played the coordination game in the reference group. We find that Democrats’ beliefs about Republican norms are consistent with actual Republican norms. Similarly, Republicans’ beliefs about Democrat norms align with actual Democrat norms. Thus, on the issue of redistribution through taxation, we show that liberals do understand conservative values and vice versa.
Chang, D. and E. L. Krupka, “Social Norms in Public Goods Games.”
In our study, we use a novel norms elicitation method from Krupka and Weber (2013) that allows us to separately and exogenously identify normative prescriptions for the social identity in that particular situation. Our study is a 2 (identity: Reserve Officers’ Training Corps (ROTC) students vs. civilian students) x 2 (situation: positive frame x negative frame) between-subject design. Participants are primed with either their ROTC identity or their student identity then randomly assigned into either the positive framed public goods game or the negative framed public goods game from Andreoni (1995). The action space and the payoffs associated with these actions are held constant across these two frames, however, in Andreoni (1995) and in subsequent variations of the paper, behavioral differences are observed (e.g. Park, 2000). Using the norms elicitation method, we directly elicit empirical measures of the social norms for the ROTC students and for the civilian students in the positive frame and in the negative frame. By comparing the behavior within identity across the positive and negative frames we hold constant any changes in social preferences that might accompany a change in which identity is made salient. Thus, we may test for differences that in the normative prescriptions associated with the frame change. Furthermore, by combining these measures with the behavioral measures obtained earlier in the study, we test for evidence of the social norms mechanism in social identity driven choice.