Impacts of workplace verbal aggression classified via text mining on workers’ mental health

Abstract

Abstract Background Exposure to workplace aggression adversely affects workers’ health; however, little is known regarding the impact of specific types of verbal content. Aims We aimed to examine the relationship between exposure to several types of aggressive words at work and the victim’s depressive symptoms and sleep disturbance using text mining. Methods We conducted a longitudinal survey with 800 workers in wholesale and retail companies; of which, 500 responded to the follow-up survey. The Centre for Epidemiologic Studies—Depression Scale and Pittsburgh Sleep Quality Index were filled out by the participants, and their responses were analysed by logistic regression to evaluate the risk of depression or sleep problems. We collected exact aggressive words encountered at work over the past year as a dependent variable and classified it into four types using text mining, such as words criticizing one’s performance. Results The follow-up rate was 63%. Exposure to words threatening one’s life showed a significant relationship with the risk of depression (odds ratio [OR] = 13.94, 95% confidence interval [CI] = 1.76–110.56). The exposure to words criticizing one’s job performance is significantly related to the risk of sleep disturbance (OR = 5.56, 95% CI = 2.08–14.88). Conclusions These findings suggest that different contents of verbal aggression can have different impacts on workers’ health. This indicates that not only overtly threatening and abusive language but also words related to one’s performance can be a risk factor for workers, depending on how they are delivered. To mitigate the adverse effects, promoting effective communication and cultivating psychological detachment from work may be beneficial.