Anonymity in social media platforms keeps users hidden behind a keyboard. This absolves users of responsibility, allowing them to engage in online rage, hate speech, and other text-based toxicity that harms online well-being. Recent research in the field of Digital Emotion Regulation (DER) has revealed that indulgence in online toxicity can be a result of ineffective emotional regulation (ER). This, we believe, can be reduced by educating users about the consequences of their actions. Prior DER research has primarily focused on exploring digital emotion regulation practises, identifying emotion regulation using multimodal sensors, and encouraging users to act responsibly in online conversations. While these studies provide valuable insights into how users consciously utilise digital media for emotion regulation, they do not capture the contextual dynamics of emotion regulation online. Through interaction design, this work provides an intervention for the delivery of ER support. It introduces a novel technique for identifying the need for emotional regulation in online conversations and delivering information to users in a way that integrates didactic learning into their daily life. By fostering self-reflection in periods of intensified emotional expression, we present a graph-based framework for on-the-spot emotion regulation support in online conversations. Our findings suggest that using this model in a conversation can help identify its influential threads/nodes to locate where toxicity is concentrated and help reduce it by up to 12\%. This is the first study in the field of DER that focuses on learning transfer by inducing self-reflection and implicit emotion regulation.

翻译：社交媒体平台的匿名性使用户隐藏在键盘后，这免除了用户的责任，使其能够在线上发泄愤怒、发表仇恨言论以及实施其他损害网络健康的基于文本的有毒行为。数字情绪调节领域的近期研究表明，沉溺于网络毒性行为可能是情绪调节无效的结果。我们相信，通过教育用户了解其行为的后果，可以减少这种现象。既往数字情绪调节研究主要聚焦于探索数字情绪调节实践、利用多模态传感器识别情绪调节，以及鼓励用户在在线对话中负责任地行事。尽管这些研究对用户如何有意识地利用数字媒体进行情绪调节提供了宝贵见解，但未能捕捉在线情绪调节的语境动态。本研究通过交互设计为情绪调节支持的传递提供了一种干预手段，引入了一种新方法：识别在线对话中情绪调节的需求，并以将说教式学习融入日常生活的方式向用户传递信息。通过在情绪表达加剧时期培养自我反思，我们提出了一个基于图的框架，用于在在线对话中提供即时情绪调节支持。研究结果表明，在对话中运用该模型有助于识别影响力较大的线程/节点，以定位毒性集中区域，并帮助将毒性降低高达12%。这是数字情绪调节领域首项专注于通过诱导自我反思和内隐情绪调节实现学习迁移的研究。