Workplace Bullying: Modelling Construct Validity in an Australian Public Sector Workforce

Abstract

Background: Measuring workplace bullying is a challenge as different workforces have different issues. Objectives: The aim of this study was to evaluate the construct validity and reliability of a Workplace Bullying Inventory (WBI) that was originally derived and validated in a nursing workforce in a broader sample of the Australian public sector workforce. Method: A two-stage procedure was employed with a large sample of volunteer public sector workforce union members (n=1,508), randomly split in two halves. The first sample (n=754) was analysed using exploratory factor analysis (EFA) and the second sample (n=754) through confirmatory factor analysis (CFA) using MPlus. The WBI consisted of 20 items related to bullying behaviours; 17 of which we had previously validated for use in an Australian nursing workforce. An additional 3 items were included from other workforce bullying scales in order to enhance the generalizability of the scale to a wider workforce. Results: The final model, identified in sample 1 and validated in sample 2 (CFA), best represented workplace bullying as three reliable first-order factors, together loading 13 items; Social Intimidation with 6 items (α = 0.88), Personal Attack with 4 (α = 0.84) and Attack through Work Tasks with 3 items (α= 0.85). Discussion: The modelling demonstrates that workplace bullying is a complex construct, consisting of multiple factors. The original measure derived from the nursing workforce demonstrates broad applicability in different public sector workforce contexts. Differences in the factor structure of the WB construct in this study when compared to the earlier nursing study suggest subtle differences may exist in the nature of bullying across workgroup settings. Future work in theory development should investigate further the complexities of the construct in order to learn how to manage and ultimately prevent WB behaviours.