Can Social polls be an accurate source of public opinion?: Correcting Biases of Twitter polls using post-stratification methods

• In Event: Public and Emerging Technologies, and Philosophies toward AI

Brief Overview

We examined Twitter polls from two US Presidential elections, uncovering biases potentially from skewed samples and inauthentic votes. Using demographic and political post-stratification, we show that social polls can measure public opinion.

Authors

• Stephen Scarano, Michigan State University

• Przemyslaw Grabowicz, University of Massachusetts-Amherst

• JungHwan Yang, University of Illinois Urbana-Champaign

• Mattia Samory, Sapienza Università di Roma

• Vijayalakshmi Vasudevan, University of Massachusetts-Amherst