Repository logo
Log in(current)
Repository logoMIT Open ScholarshipDSpace@MIT
  1. Home
  2. MIT Libraries
  3. MIT Theses
  4. Doctoral Theses
  5. Essays on Ecological Rationality and Political Beliefs

Essays on Ecological Rationality and Political Beliefs

Thumbnail Image
Download
Name

Orchinik-orchinik-Management-2026-thesis 1.pdf

Size

25.91 MB

Format

Adobe PDF

Checksum (MD5)

d142e6a419be71c5647c53653ce3a864

Author(s)
Orchinik, Reed
Advisor(s)
Bhui, Rahul
Date Issued
May 2026
Publisher
Massachusetts Institute of Technology
Abstract
Despite unprecedented access to information, much of it accurate, people hold misinformed and polarized beliefs. A common explanation is that human cognition - particularly in the context of politics - is plagued by biases that leave people vulnerable to exploitation by social media, advertisements, and political propaganda. I argue instead that people possess a suite of ecologically rational cognitive tools, including intuitions and mental models, that are tailored to and help them navigate complex information environments under constraints. These adaptations often work well but simultaneously create predictable vulnerabilities. Chapters 1 and 2 focus on intuitions, examining how they adapt to the statistical structure of information environments. From this perspective, I argue that belief in misinformation is not solely a cognitive failing, but a rational byproduct of cognition adapted to environments that are usually, but not always, high-quality. As the quality of the environment deteriorates, skepticism can become intuitive. Chapter 3 turns to the information environment itself, examining the political narratives that aid consumers in interpreting a fast-paced news environment while framing political events for them. Chapter 1 - joint with Cameron Martel, David Rand, and Rahul Bhui - demonstrates that intuitions, which are generally understood to make people gullible, are calibrated to the base rate of accurate content. Across a large-scale preregistered survey experiment and an incentivized replication, I test whether intuitions adapt to the base rate of true versus false content. Participants viewed news feeds composed primarily of either true or false headlines. Individuals make faster and more frequent errors when encountering the less frequent headline type, and fewer errors with the more common type. Computational modeling confirms these effects are driven by intuitive responses that function like Bayesian priors about the veracity of content. Cross-culturally, the magnitude of intuitive truth bias varies with the quality of the national information environment. Chapter 2 extends this adaptive interpretation of intuitions to the illusory truth effect - the finding that repetition increases belief. A formal model shows that believing repeated statements improves accuracy when sources are credible but occasionally err, unifying key findings in the literature while generating a novel prediction: the effect should attenuate when sources lack credibility. Two large, preregistered experiments confirm this prediction. In the first, the illusory truth effect is substantially smaller, and for most participants absent entirely, when the source is likely to spread falsehoods. In the second, the effect disappears for statements attributed to a distrusted presidential candidate. These studies provide evidence that the adaptive use of repetition operates through intuition rather than deliberation. Chapter 3 is joint work with Baird Howland, Duncan Watts, and David Rothschild. This chapter explores narratives - causal stories that link recent events to their underlying causes. These narratives provide quick and portable explanations that news consumers use to interpret politics. Here, I focus on the information environment itself, examining how news outlets construct political narratives in the two years surrounding the 2024 U.S. Presidential Election. Using approximately 120,000 articles from six prominent U.S. news outlets published between January 2024 and October 2025, I employ a natural language processing pipeline to identify a set of focal narratives and characterize them along six dimensions. During this period, right-leaning outlets produce more narratives that are more frequently repeated, span a wider range of events, and maintain greater internal coherence. In the lead-up to the election and the transition between administrations, the overall landscape skews pro-Republican, driven largely by an anti-incumbency dynamic in which negative coverage of the sitting administration benefits the opposing party. This pattern of anti-incumbency persists as outlets shift their targets to the Trump administration during Trump’s first 9 months in office. A recurring exception is the "strongman" narrative: ostensibly negative portrayals of Donald Trump that nonetheless carry a pro-Republican lean by emphasizing political effectiveness and institutional dominance.
MIT Department
Sloan School of Management
Terms of Use
In Copyright - Educational Use Permitted
https://rightsstatements.org/page/InC-EDU/1.0/
Copyright retained by author(s)
Persistent DSpace Link
https://dspace.mit.edu/handle/1721.1/174596
Repository logo
PrivacyPermissionsAccessibilityContact us
Repository logo
Notify us about copyright concerns.