Interactive crowdsourcing to fact-check politicians.

Area of Science:

• Social Psychology
• Political Science
• Information Science

Background:

• Political leaders' discourse frequently contains misinformation, necessitating public guidance.
• Fact-checking agencies struggle with scalability in verifying political speech.
• Crowdsourcing shows potential for scalable news fact-checking, but its efficacy for political speech remains uncertain.

Purpose of the Study:

• To investigate the impact of social influence on the accuracy of collective judgments regarding the veracity of political speech.
• To determine if crowdsourced judgments can effectively verify politicians' statements.

Main Methods:

• Two experiments were conducted with 180 and 240 participants, respectively.
• Participants initially judged the veracity of political phrases.
• Revised judgments were collected after exposure to homogeneous or heterogeneous social information.

Main Results:

• Heterogeneous social influence significantly increased the accuracy of participants' judgments compared to a control group.
• Homogeneous social influence did not yield a significant improvement in accuracy.
• Collective judgments on political speech veracity are susceptible to social influence dynamics.

Conclusions:

• Social influence, particularly from diverse perspectives, can enhance the accuracy of crowdsourced political speech verification.
• Interactive crowdsourcing strategies incorporating heterogeneous influence show promise for fact-checking agencies.
• Findings contribute to understanding collective intelligence in the context of political misinformation.