Abusive Supervision, Job Stress, and Cyberloafing in Hospital Settings: The Mediating Role of Job Stress and Moderating Role of Self-Control

Abstract

Cyberloafing the non-work-related personal use of organizational internet resources during working hours poses a particularly consequential threat in healthcare settings, where service quality and patient safety depend critically on sustained employee attention. This study examines an integrated moderated-mediation model in which abusive supervision predicts cyberloafing both directly and through the intermediary mechanism of job stress, with self-control serving as a boundary-condition moderator of both stress-generating and stress-translating pathways. The study was conducted at RSUD Linggarjati and RSU El-Syifa in Kuningan Regency, West Java, Indonesia. A quantitative causal-verificative design was employed, with data gathered from 218 respondents (stratified proportional random sampling; Slovin formula, e = 5%) using closed-ended Likert-scale questionnaires. Structural Equation Modeling–Partial Least Squares (SEM-PLS) via SmartPLS 4 was used for hypothesis testing. Outer model evaluation confirmed satisfactory validity (all loadings ≥ 0.70; AVE ≥ 0.50; HTMT < 0.90) and reliability (CR ≥ 0.70; α ≥ 0.70). Inner model results reveal strong predictive performance: R² (cyberloafing) = 0.919; R² (job stress) = 0.512; SRMR = 0.045–0.046. Abusive supervision exerts significant positive direct effects on both cyberloafing (β = 0.274; p < 0.05) and job stress (β = 0.715; p < 0.05). Job stress significantly predicts cyberloafing (β = 0.861; p < 0.05) and partially mediates the abusive supervision–cyberloafing relationship. Self-control significantly moderates both the abusive supervision–job stress pathway and the job stress–cyberloafing pathway (negative interaction coefficients; both p < 0.05), buffering the escalation of stress under abusive leadership and curtailing the stress-to-cyberloafing translation. PLS-Predict confirms strong out-of-sample predictive relevance for all endogenous indicators. Findings advance theory at the intersection of Social Exchange Theory.