Investigating the impact of decent work perception on employee voice behavior: evidence from text mining

The purpose of this study is to explore how five specific sub-items of employees’ perception of decent work affect their voice behaviors, particularly prohibitive voice behaviors. This research aims to delve deeper into this niche area and investigate the boundary conditions of this relationship from the perspective of CEOs. 203,197 valid employee reviews were gathered from Glassdoor using text mining and sentiment analysis techniques. To test the hypotheses, negative binomial regression analysis was employed on the results. The results reveal that the perception of decent work positively and significantly affects both employee voice behaviors and their prohibitive voice behaviors. Additionally, CEO approval significantly moderates the relationship between the perception of decent work and voice behaviors. Interestingly, CEO approval reinforces the effect of decent work on employee voice behaviors. However, the moderating effect of CEO approval on prohibitive voice behaviors is somewhat reversed. This study has used the application of text mining techniques within the context of employees’ decent work perception, delving into the nuanced relationship between decent work and voice behaviors. This approach has expanded the research landscape within this specialized field.