Search

This project explores how entrenched attitudes and conventions change over time, and whether this change can be facilitated by exogenous interventions. Specifically, we investigate whether there are certain distributions of beliefs or preferences where a policy maker cannot facilitate endogenous social change by incentivizing a subset of the population. In recent years, social norm change has been added to the policy toolkit, alongside nudges and other choice architecture interventions, as a means for engendering widespread shifts in beliefs or behaviors in a society without the introduction of legal mandates or regulation. Social norm interventions are particularly effective when a small intervention has outsized effects on a society through endogenous norm change dynamics or “spillovers” to the rest of the population. However, a small intervention is only guaranteed to induce widespread changes in attitudes of behaviors under certain distributions of attachment to the attitude or behavior in the population. Theoretical models show that heterogeneity in this attachment has important implications for how and whether a shift in behaviors will spread through a population, and who should be targeted by an intervention. We test the empirical predictions of these models in 1. a virtual coordination game in which we vary the attachment to different arbitrary behaviors in the game using monetary incentives, and 2. a coordination game where we induce attachment to the behaviors using labels that are either aligned or in conflict with the attitudes and beliefs of our respondents. Specifically, we use political labels in the highly polarized context of the U.S. to create variation in the attachment to otherwise arbitrary actions in a coordination game. We then intervene on a subset of our population (“target group”), altering their monetary incentives to be in conflict with their political preferences. This allows us to test whether degree of attachment to an opinion is an important moderator of endogenous social change using preferences that closely resemble those of the real world contexts in which these interventions are being considered. Finally, we treat the election as a natural experiment that changes the values associated with different political labels and look at how this affects choices in the coordination game.

Data and code to replicate the analyses in "Group Identities Can Undermine Social Tipping after Intervention"

The popular practice of 'leading by the successful' is viewed as a hallmark of motivational leadership. A central rationale for leaders to make successful team members salient is that it may induce social learning, where followers strive to adopt a favorable behavior. The reliance of a leader on such success-biased social learning presumes that imitation by followers occurs only to the extent as outstanding success was caused by a superior ability or knowledge of the respective peer. In this article, we conduct a laboratory experiment to study whether imitation of the successful may occur even if imitation necessarily fails to be an effective way of improving one's performance. The experimental approach establishes the necessary control to assure that success-biased learning cannot systematically improve the decisions made, and allows us to isolate the behavior of the followers from possible feedback effects of the leader. The data show that a substantial amount of imitation occurs, which in our setting leads to a sizeable and persistent increase of the average risk taken in the teams. Our finding thus indicates a limitation of the practice to lead with the successful.

EN: Unity and flexibility in the future of the European Union: The challenge of enhanced cooperation. Pp. 47-72