Proxy measure of safety

Safety culture (and/or climate) measures may themselves be regarded as proxy measures of safety in the sense outlined above. However, few studies have found these measures to be strongly related to hard risk outcomes such as injuries and accidents (Guldemnund 2000 The Health Foundation 2011). There are several reasons for this apparent lack of correlation. Safety climate or culture is, by definition, shared within a social unit (a work group), but such units are usually ill defined and small. Safety climate or culmre is multi-faceted, and each facet is a constract, as described in the previous sections of this chapter, based on a few items from a questionnaire. Although the reliability and the intra-class correlation for the constructs can be acceptable, repeated measurements are typically infeasible, and when the questionnaire has been applied repeatedly, its responsiveness (the ability of the constract to reliable measure changes over time (de Vet et al. 2011)) is usually not reported but can be expected to be low. At the same time, since... 

---