Development of a Troll Behavior Perception Scale in Artificial Intelligence Tools

This study developed a valid and reliable scale to measure perceptions of troll behavior in artificial intelligence tools. As digital technologies and AI-based applications become increasingly widespread across various fields, including education, the need to examine the potential social and psychological effects of online trolling has grown. In educational settings, ensuring a safe, ethical, and healthy digital interaction environment requires that such behaviors be measurable. The study used convenient sampling methods to reach participants from 300 different age groups and 33.7% of this group were male and 66.3% were female. Most participants were between the ages of 18 and 31. The data collection instrument was a 22-item, five-factor, 5-point Likert scale. The factors included aggressive and hostile behaviors, data manipulation, social provocation, anonymity and multiple identities, and emotional manipulation. Exploratory Factor Analysis (EFA) revealed a five-factor structure aligned with these dimensions, explaining 74.135% of the total variance and supporting the construct validity of the scale. Confirmatory Factor Analysis (CFA) further validated the measurement model by demonstrating statistically significant relationships between subdimensions and scale items. Overall, the findings demonstrate that the scale is a valid and reliable tool for assessing the influence of AI-related tools on perceptions of troll behavior, particularly within educational contexts.