Workplace bullying report by bystanders in Colombia: an exploratory analysis of social media data using machine learning algorithms

Abstract

Purpose: This study characterizes how bystanders report workplace bullying in Colombia through social media, exploring both descriptive and predictive dimensions of these reports.
Methodology: The study employs text mining and predictive algorithms to analyze bystanders' reports of workplace bullying. The analysis focuses on elements such as report length, frequent mentions, references to women's issues, and arguments against aggressors.
Results: Significant predictive variables included: the emotional tone of the reports, the frequent mention of the victim's gender, and the minimal use of emojis. These variables provide robust indicators for understanding how workplace bullying is reported by bystanders.
Implications: The study underscores the importance of adopting a gender-sensitive approach and considering structural and cultural factors when addressing workplace bullying. It also highlights the need for further research to systematically synthesize findings and to inform Policies and regulations aimed at effectively combating workplace bullying.
Originality: This research makes a novel contribution by focusing on workplace bullying reports from the perspective of bystanders in Colombia, an underexplored angle in the literature. The integration of social media data with text mining and predictive algorithms offers an innovative methodological approach that expands understanding of workplace bullying dynamics.