How Political Content Quality Is Distributed across Race, Gender, and Ideology

• In Event: Social Identity, Power, and Mis- and Disinformation

Abstract

While mis- and disinformation have attracted substantial attention in political science and related disciplines over the past several years, researchers quickly moved past the binary “fake/real news” paradigm that anchored early studies. Recent developments in political content classification have led to the widespread use of empirical rankings of web sites by content quality. High-quality sites consistently deliver accurate and verifiable information as transparently as is feasible, while low-quality sites are rife with distortions, manipulations, and outright falsehoods (although some content may be true). Sites ranked in the middle deliver an inconsistent mix of true and deceptive content. We combine one such scale with a unique research design that integrates survey data with personalized timeline content from Twitter to discover the demographic and psychographic characteristics (prominently including race, gender, and ideology) most associated with substantial proportions of high- and low-quality content. We further classify the Twitter accounts appearing most often in our users’ timelines by gender, race, and ideology to better understand how low- and high-quality political content is targeted and distributed.

Author

• Deen Freelon, University of Pennsylvania