A contemporary view of interpersonal aggression and cyberbullying through ICT: multilevel insights from LMX differentiation

Purpose While the rapid adoption of information communication technologies (ICT) in organizations has been linked with a higher risk of cyberbullying, research on the influence of cyberbullying on interpersonal behaviors in the workplace remains limited. By drawing on the ego-depletion theory and the leader-member exchange (LMX) theory, this research investigates how, why and when workplace cyberbullying may trigger interpersonal aggression through ICT. Design/methodology/approach The authors collected data from 259 employees and 62 supervisors working in large ICT organizations in China through a multi-wave survey. The authors performed multilevel analysis and used hierarchical linear modeling to test the proposed moderated mediation model. Findings The results revealed that workplace cyberbullying has a significant and positive influence on interpersonal aggression in the workplace via ego depletion. The authors found that differentiation in LMX processes at group level moderates the indirect relationship between workplace cyberbullying and interpersonal aggression (via ego depletion). Furthermore, the positive indirect effect of workplace cyberbullying was found to be stronger in the presence of a high LMX differentiation condition in comparison to a low LMX differentiation condition. Research limitations/implications The data were collected from Chinese ICT organizations, which may limit the generalization of this study's findings to other cultural and sectoral contexts. Originality/value This paper provides the first step in understanding how, why and when workplace cyberbullying triggers interpersonal aggression by investigating the role of ego depletion as a mediator and LMX differentiation as a boundary condition. This is the first study to empirically examine the relationships between workplace cyberbullying, ego depletion, LMX differentiation and interpersonal aggression in ICT organizations using multi-level modeling.