preference analysis of content moderation by YouTube

Abstract

This study examines content moderation of information that is not clearly illegal but generally considered undesirable, from the perspective of user preferences. Using YouTube as a case study, we focus on three subcategories of age-restricted content —child safety, harmful or dangerous activities, and vulgar language—and quantify users’ preferences regarding whether such content should remain available or be removed, and who should have authority over removal.

We begin by reviewing recent policy and legal developments related to content moderation. We then review prior research and, on this basis, formulate three research questions: (1) whether users wish to view such content or prefer that it be deleted; (2) if content is to be deleted, who should decide—the platform or the users themselves; and (3) whether those favoring permissive policies share characteristics with participants in social-media firestorms or contributors to fake news. These questions are examined through a web survey.

The analysis shows, first, that attitudes toward viewing or removing content that is deemed undesirable are sharply divided and associated with users’ sociodemographic attributes and their experiences and beliefs related to YouTube. Second, preferences over who should decide on deletion are similarly split between leaving decisions to YouTube and having users themselves make those decisions and are associated with users’ exposure to edited media such as news sites and television. Third, factors that prior research has identified as typical among individuals who engage in harassment or spread misinformation, such as gender and income, only partially overlap with those who favor more open content policies. From a policy perspective, the findings highlight the importance of fostering user opinion formation and deliberation that bring together diverse viewpoints grounded in a wide range of attributes, experiences, and beliefs.