Toxic Behaviours in Esport: A Review of Data-Collection Methods Applied in Studying Toxic In-Gaming Behaviours

Research into toxic behaviors within esports predominantly relies on self-reported accounts, such as surveys and interviews, which offer valuable insights into player experiences and coping mechanisms but are limited by recall bias and an inability to capture real-time, naturalistic behaviors. While online log data effectively maps the occurrence of toxicity and its correlation with game outcomes, it is restricted to textual forms and lacks broader contextual information. Observational methods provide some contextual understanding, but non-participatory approaches are often limited to forum activity, and participatory observations risk researcher influence.

A significant gap exists in the application of methods capable of capturing naturalistic toxic behaviors as they unfold. The analysis of 54 empirical articles reveals that less established methods are underutilized. A wide range of toxic behaviors, encompassing both "Chat actions" like hate speech and insults, and "Gameplay actions" such as AFK and intentional feeding, are investigated using these diverse, yet often limited, data collection approaches.

The document proposes that online video-based behavioral analysis holds substantial potential to address this research deficit. This novel approach, drawing parallels from offline violence research, offers a promising method for fine-grained, direct observation of natural toxic behaviors. It provides detailed contextual information and insights into the temporal properties of toxic interactions, overcoming the limitations of both direct observation and automatic log data. This method is assessed as highly valid and reliable for assessing encounters with toxic behaviors, despite potential challenges related to visual access and the authenticity of footage.

The existing body of research, as evidenced by a comprehensive bibliography, covers various aspects of toxic behavior, including linguistic analysis, the role of personality and contextual factors, and the development of detection tools. However, a shift towards more direct, real-time observational methods, particularly video-based analysis, is advocated to enhance the understanding of toxic behaviors in esports by capturing their naturalistic occurrence and intricate contextual details.