PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
January 18, 20260 citationsOpen Access

Overprecision and (Ir)rational Inattention

View Full Paper
CBCiril Bosch-RosaMBMuhammed BulutayBKBernhard Kassner

Key Points

  • The research aims to examine how overprecision influences belief updating in decision-making processes.
  • Conducted a pre-registered 2 x 2 belief-updating experiment.
  • Manipulated levels of overprecision and information costs.
  • Measured belief updating responsiveness to new information.
  • Overprecision significantly reduces the rate of belief updating.
  • Higher information costs lead to a decreased responsiveness to signals.
  • No predicted interaction found between overprecision and attention costs, though overprecise participants show stronger effects from information costs.

Abstract

This paper integrates overprecision—a form of overconfidence where individuals overestimate the accuracy of their beliefs—into a canonical rational inattention model. We show that overprecision distorts belief updating directly by biasing the perceived value of new information and indirectly by amplifying the impact of attention costs. We test the model’s predictions in a pre-registered 2 x 2 belief-updating experiment that manipulates overprecision and information costs. The results confirm that overprecision reduces updating and that higher information costs lead to lower responsiveness to signals. While our pre-registered analysis finds no support for the predicted interaction between overprecision and attention costs, a more granular specification shows that the effect of information costs on belief updating is stronger for overprecise participants. These findings suggest that what appears as rational inattention partly reflects irrational inattention arising from misperceived prior accuracy, underscoring the need to distinguish informational frictions from cognitive biases when modeling attention.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Bosch-Rosa et al. (2026) studied this question.

synapsesocial.com/papers/696c785beb60fb80d13968d5https://doi.org/10.48462/opus4-6028
Ask AI
Helpful
Bookmark
Share
View Full Paper