What Do the Sailors Say? Using Natural Language Processing to Interpret Open-Ended Responses about Safety Issues
Abstract
Occupational safety is a critical component of organizational success, especially in industries with high-risk working conditions. Maximizing an organizations occupational safety outcomes requires routinely assessing its safety climate to help identify emerging issues as well as opportunities for improvement. Naval operations present particularly complex environments since working and living spaces are intertwined and personnel serve in prolonged high-risk operations. Although many safety climate assessments probe this environment with structured surveys, there is the opportunity for the survey measure itself to inadvertently bias answers through pre-determined responses. To address this issue, open-ended questions allow personnel to report safety-related concerns without inducing surveyor bias. The current study utilized natural language processing and topic modeling to explore open-ended safety answers from a large sample of naval personnel. Answers indicated that personnel and readiness were among the most frequently reported safety-related problems along with scheduled sleep and watchstanding. Moreover, senior leaders offered notably different comments about the nature of best and worst safety-related practices compared to junior personnel. This outcome demonstrates some common elements in the most prominent safety concerns in maritime operations while also noting some systemic differences in perceptions that are influenced by position within the leadership hierarchy.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 11, 2023
- Accession Number
- AD1207890
Entities
People
- Adam T. Biggs
- Dale W Russell
- Ryan Lance
- Todd R. Seech