Belief polarization in a complex world: A learning theory perspective

Authors:
Publisher website
Google Scholar
Abstract
Significance: Differences in beliefs within a society are a prevalent human phenomenon. A standard explanation for polarized beliefs relies on “echo chambers” that expose people to different sources of information. However, there is ample evidence that people sustain different beliefs even when faced with the same information, and they interpret that information differently—facts often attributed to a confirmatory bias. We suggest models of how humans form beliefs based on machine learning theory. These models show how stark differences in beliefs can arise and endure due to human limitations in interpreting complex information. Our framework illuminates inherent challenges and potential ways of overcoming polarization.
Funding Information
  • National Science Foundation (SES-1629446)
  • National Science Foundation (SES-2018554)

This publication has 36 references indexed in Scilit: