The sounds of safety silence:interventions and temporal patterns unmute unique safety voice content in speech

Research shows that withholding safety concerns on encountering hazards – safety silence – is a critical contributor to accidents. Studies therefore aim to prevent accidental harm through interventions for reducing safety silence. Yet, the behaviour remains poorly understood, obstructing effective safety management: it is unclear to what extent safety silence involves muted safety voice (the partial withholding of safety concerns), and how muted safety voice can be recognised in speech, may be measured based on the degrees and types of safety voice (speaking up about safety), progresses over time, and may be optimally reduced. To improve safety management, this study proposes a conceptual model for the manifestation of safety silence and muted safety voice using a laboratory experiment (N = 404) to evaluate the implications for the effectiveness of three interventions (salient hazards, clear responsibilities, encouragements) across stages of a hazard. Results indicated that safety silence and muted safety voice are measurable in terms of the degree to which concerned people engage in five types of safety voice at different points in time, and we revealed this is important for safety management: interventions only unmute safety voice at unique hazard stages and for knowledge-based speech when people are concerned. This indicates that safety silence and muted safety voice are situated and can be recognised in nuanced speech, with interventions being most effective when timed appropriately and people have safety concerns to speak up about.