Determining the optimal cut-off scores for the workplace bullying inventory

Abstract

Background Over the past two decades, there has been considerable research into workplace bullying. One area that remains poorly developed is a tool with the capacity to accurately differentiate between exposed and unexposed employees. Aim To determine optimal cut-off scores for the Workplace Bullying Inventory (WBI) that accurately classify cases of exposure to workplace bullying. Discussion Secondary analysis of data collected from Australian public sector employees (n=2,197) was conducted. Receiver operator characteristic (ROC) curve analysis was used with a minimum sensitivity of 80%, to determine those scores on the WBI that corresponded with the highest accuracy of the tool to distinguish cases from non-cases. The results suggest using a cut score of 29 from the total score on the WBI (possible range: 18-90). When compared to a sum-score from a single dichotomous self-report variable, the cut-off score estimated a more conservative bullying rate. The single-item rate was potentially inflated by misconceptions about what constitutes bullying in the workplace. Conclusion Employing validated cut-off points for exposure provides an objective threshold for establishing exposure to workplace bullying. The results of the analysis provide a more rigorous approach to quantifying exposure to workplace bullying, in a tool that has been designed and tested in the nursing workforce. This is the first such tool with empirically-derived, discriminant accuracy. Implications for practice It is common for nurse researchers to employ sum-scores from single items to identify exposure to workplace bullying. By providing reliable cut-off points for exposure, this study offers standardised, diagnostic accuracy for researchers, clinicians and managers.