When Does an Individual Accept Misinformation? An Extended Investigation Through Cognitive Modeling

Area of Science:

• Cognitive Psychology
• Social Media Studies
• Computational Social Science

Background:

• Misinformation and disinformation spread rapidly on social media platforms like Twitter.
• Existing models often focus on average fake news judgment, neglecting individual decision-making processes.
• Predicting individual responses to fake news requires analyzing raw user data and adapting theoretical models.

Purpose of the Study:

• To comparatively assess the predictive accuracy of various socio-psychological and cognitive models for individual fake news judgment.
• To determine which models best predict an individual's decision to accept or reject misinformation before the decision is made.
• To refine existing modeling approaches using large-scale datasets.

Main Methods:

• Utilized three large, previously collected datasets totaling 3794 participants.
• Implemented and adapted classical reasoning, sentiment analysis, and heuristic modeling approaches.
• Compared the predictive performance of various models, including "motivated reasoning" and hybrid combinations.

Main Results:

• Classical reasoning, sentiment analysis, and heuristic models demonstrated the highest predictive accuracy for individual "Accept" or "Reject" responses.
• A model derived from Jay Van Bavel's research led in predictive performance.
• Hybrid models, combining multiple individual models, significantly improved prediction performance.

Conclusions:

• Individual decision-making in response to misinformation can be effectively predicted by specific modeling approaches.
• A combination of classical, sentiment, and heuristic models offers a robust framework for understanding fake news acceptance.
• The findings suggest an "adaptive toolbox" of models for analyzing individual responses to online information.