Loading Events
  • This event has passed.
Other Events

SES & IDPS Dissertation Defense

August 21, 2023 @ 10:30 am - 12:30 pm

Amir Tohidi (IDSS)


Habit Formation and Political Persuasion: A Behavioral and Statistical Approach


This talk explores the complex dynamics of human behavior across diverse contexts, integrating perspectives from behavioral science and statistical analysis. The central focus revolves around the analysis of repetitive behavior in  various scenarios. Through a multidisciplinary approach, we delve into the intricacies of habit formation and political persuasion, unraveling their underlying mechanisms and societal implications.

The first project investigates the influence of habits on the in-store shopping experience. By utilizing panel data on household purchases and leveraging store closures as a disruptive event, we examine how these closures prompt individuals to alter their purchasing patterns. Employing a difference-in-differences framework, we estimate the causal impact of habits on brand loyalty. Our findings reveal a significant role of habits, with households exhibiting stronger habits experiencing a temporary disruption in their shopping routines following store closures. Over time, these households appear to develop new habits in different stores, resulting in lasting changes in preferred brands. This suggests that the formation of shopping habits can lead to suboptimal consumer behavior. These insights have practical implications for businesses, including pricing strategies, advertising approaches, and product placement within stores.

The second project introduces an innovative methodology for quantifying habitual behavior in the context of social media usage. Interactions with social media platforms often yield psychological rewards, fostering the development of habitual behaviors driven by cue-response associations. By leveraging entropy as an implicit measure of behavioral regularity, this study aims to uncover the intricate relationship between habit formation and digital routines. Through empirical analyses, we establish the validity of the entropy metric, demonstrating its effectiveness in capturing distinct behavioral patterns beyond mere frequency. Our findings contribute to the theoretical understanding of habitual behavior and offer practical insights for managing digital habits. Ultimately, this work advances our comprehension of how habits manifest in the digital realm and provides a robust tool for predicting long-term user behavior.

The final chapter delves into the intricate interplay between individuals’ beliefs and their ability to anticipate the persuasive impact of climate change news articles. The central aim is to determine whether climate change deniers or believers possess varying capacities to predict the persuasive consequences of articles emphasizing climate change severity. Through a series of surveys with US participants, we gather predictions about the impact of such articles on climate change deniers. Surprisingly, findings reveal discordant predictions: deniers anticipate a backfire effect among peers, while climate believers foresee negligible effects. We rigorously test these predictions with a randomized survey experiment involving deniers, uncovering an unexpected positive opinion shift towards climate change after article exposure. Notably, this effect does not translate into discernible changes in stated or revealed support for climate change actions. In the context of the pressing climate challenge, our study offers insights to inform targeted communication and interventions that foster consensus and meaningful action.


Ali Jadbabaie (supervisor), Dean Eckles (supervisor), Adam Berinsky, Drazen Prelec

Event Information

Hybrid event. To attend virtually, please contact the IDSS Academic Office (idss_academic_office@mit.edu) for connection information.

MIT Institute for Data, Systems, and Society
Massachusetts Institute of Technology
77 Massachusetts Avenue
Cambridge, MA 02139-4307