When and how vicarious abusive supervision leads to bystanders' supervisor-directed deviance A moderated-mediation model

Purpose The purpose of this paper is to examine bystanders' supervisor-directed deviance to vicarious abusive supervision by supervisor-directed attribution. Furthermore, this study developed a moderated-mediation model to explore how LMX between bystander and his/her supervisor moderate the relationship between vicarious abusive supervision and the supervisor-directed attribution, which subsequently influences bystanders' supervisor-directed deviance. Design/methodology/approach The paper tested the model using a sample of 336 workers using a two-wave survey. A moderated-mediation analysis was conducted with bootstrapping procedure to test the first stage moderated-mediation model in this study. Findings The results showed that LMX (between bystander and his/her supervisor) weakens the indirect relationship between vicarious abusive supervision and supervisor-directed deviance by bystanders' supervisor-directed attribution. Originality/value These results clarify the nature and consequences of LMX (dyadic relationships of bystanders-supervisor) for bystanders' attribution process, and explain underlying attributional perceptions and reactions to vicarious abusive supervision. This study provides a more nuanced understanding of when and how vicarious abusive supervision leads to bystanders' supervisor-directed deviance.