Belief-consistent information is most shared despite being the least surprising.

Area of Science:

• Information Science
• Social Psychology
• Computational Social Science

Background:

• Classical information theory posits that information value is proportional to novelty or surprise.
• Prior research suggested false news spreads faster due to higher novelty, aligning with information theory.
• However, semantic meaning, such as consistency with prior beliefs, may also impact information value and sharing behavior.

Purpose of the Study:

• To investigate the roles of information novelty and semantic meaning (belief consistency) in online sharing behavior.
• To challenge the sole reliance on novelty as a predictor of information value and virality.
• To explore how pre-existing beliefs influence the perceived value and dissemination of information.

Main Methods:

• Analysis of sharing behavior on Twitter to examine initial relationships between novelty, belief consistency, and sharing.
• Two controlled experiments were conducted to manipulate information novelty and belief consistency.
• Measured sharing behavior and perceived surprise in response to belief-consistent versus belief-inconsistent information.

Main Results:

• Sharing behavior on Twitter showed separate associations with novelty and belief consistency.
• In controlled experiments, belief-consistent information was shared more frequently than belief-inconsistent information.
• Manipulated novelty did not significantly predict sharing behavior or perceived surprise; belief-consistent information was least surprising but shared most.

Conclusions:

• Classical information theoretic predictions on perceived value and sharing require refinement.
• Semantic meaning, specifically belief consistency, plays a crucial role in information sharing, independent of novelty.
• Future models of online information diffusion should incorporate the impact of pre-existing beliefs and semantic congruence.