Bifactor modeling assessment of a safety climate index using item response theory

Abstract

Effective workplace operations depend upon a collection of norms and practices, which encapsulate an overall safety climate. Although the safety climate concept has been well-defined, a continuing challenge has been developing practical and salient measurement tools. Comprehensive safety climate measurements often require complex, multi-faceted structures that can be difficult and time-consuming to administer in operational environments. The current study explored one operationally-proven survey tool with the goal of reducing the factor structure and producing a more readily useful measurement of safety climate. Statistical techniques utilized Item Response Theory and bifactor modeling to produce a Safety Climate Index (SCI). Furthermore, analyses demonstrated the practical and predictive value of this measure by comparing SCI outcomes to real safety issues experienced onboard more than 100 US Navy ships. The results suggest that the SCI could serve as a reliable and operationally practical assessment of safety climate for high-risk organizations.