Research purpose: The purpose of the study is to investigate the extent to which supervisors are universally perceived as abusive across those they supervise. I propose social learning theory and social information processing theory as theoretical bases for understanding collective impression formation among subordinates reporting to the same supervisor. The study, therefore builds on a growing body of abusive supervision literature by analysing intraclass correlations between subordinates' and their co-workers' perceptions of the same supervisor. Research motivation: Studies that examine whether or not subordinates of the same supervisor have similar perceptions of abuse are in short supply. Therefore, this study examines the possibility of objective impression formation with regards to abusive supervision so as to answer the question: Do subordinates and their co-workers mutually perceive the extent to which a supervisor's behaviour is abusive? Research design, approach and method: This study follows a cross-sectional approach to investigate the extent to which subordinates of the same workgroup mutually perceive their supervisor's behaviour as abusive. Purposive sampling was employed to recruit full-time employed Master's of Business Administration (MBA) students from six respected universities in the United States (US). Purposive sampling was further aided by snowball sampling where each subordinate was asked to get two of their own co-workers involved in the study. A total of 1,029 surveys were distributed and 693 completed surveys were returned. The final sample consisted of 210 sets of surveys where responses were received from the focal subordinate and two of his or her co-workers. An intraclass correlation coefficient (ICC) analysis was conducted to determine the strength of intra-group agreement regarding abusive supervision perceptions. Main findings: The results indicate that there is agreement between subordinates' and their co-workers' perceptions of abusive supervision as no significant differences were found between these two groups' assessments of the same supervisor's behaviour. The null hypothesis was accepted. A non-hypothesized finding is that dyadic tenure appears to influence the extent to which supervisory abuse is observed. That is, subordinates and their co-workers may view the same supervisor in the same light, the longer the duration of the supervisor-subordinate relationship. Limitations: The results should be interpreted, bearing in mind that there is limited literature available on abusive supervision at the group level. Furthermore, the results should be considered with caution as the perceptions of abusive supervision were only examined at one point in time, the use of snowball sampling method may be associated with the possibility of sampling bias, and that dyadic tenure was measured with a categorical response (i.e., not treated as a continuous variable). Finally, the results may not be generalisable to the South African context. Future research: It is suggested that future studies should investigate abusive supervision as a group-level phenomenon as few such studies currently exist. Additionally, future studies should examine the extent to which social learning and social information processing approaches contribute to the establishment of mutual perceptions about supervisory abuse. Researchers may also investigate the occurrence of abusive supervision at the group level through the lens of an alternative theoretical framework such as social identity theory.
Dissertation (MCom)--University of Pretoria, 2017.