Scholarlink

<p>With over 4 billion people using social media platforms daily, decisions about what content to allow or moderate have profound implications for global public discourse, democratic deliberation, and users’ psychological well-being. This study examines cross-national variation in content moderation preferences by exposing participants in 10 countries to a standardized example of hate speech: a hateful post modeled on text that X (formerly Twitter) had cited in its content moderation guidelines. Across contexts, exposure to such content triggers widespread support for intervention, with minimal differences based on the experimentally varied identity of the target. However, significant variation emerges across countries. The United States stands out for its comparatively lower support for moderation of hate speech, while countries like France, Brazil, and South Africa exhibit much stronger demand. Within most countries, political ideology predicts support: individuals on the right are less likely to endorse removal or suspension than those on the left. Crucially, values around freedom of speech versus harm prevention consistently shape preferences across all contexts. Those who prioritize prevention from harm are markedly more likely to support hard forms of moderation (e.g., account suspension and content take-downs). These findings highlight that moderation preferences are shaped less by the specific target than by individuals’ deeper commitments regarding free expression and harm prevention.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Ideology and free speech values predict content moderation preferences: cross-national evidence across targets of hate speech

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月17日

查看全文
  • Yannis Theocharis,
  • Spyros Kosmidis,
  • Jan Zilinsky,
  • Friederike Quint

摘要

With over 4 billion people using social media platforms daily, decisions about what content to allow or moderate have profound implications for global public discourse, democratic deliberation, and users’ psychological well-being. This study examines cross-national variation in content moderation preferences by exposing participants in 10 countries to a standardized example of hate speech: a hateful post modeled on text that X (formerly Twitter) had cited in its content moderation guidelines. Across contexts, exposure to such content triggers widespread support for intervention, with minimal differences based on the experimentally varied identity of the target. However, significant variation emerges across countries. The United States stands out for its comparatively lower support for moderation of hate speech, while countries like France, Brazil, and South Africa exhibit much stronger demand. Within most countries, political ideology predicts support: individuals on the right are less likely to endorse removal or suspension than those on the left. Crucially, values around freedom of speech versus harm prevention consistently shape preferences across all contexts. Those who prioritize prevention from harm are markedly more likely to support hard forms of moderation (e.g., account suspension and content take-downs). These findings highlight that moderation preferences are shaped less by the specific target than by individuals’ deeper commitments regarding free expression and harm prevention.