While organizations are increasingly encouraged to evaluate safety culture in the context of measurement of leading safety indicators within a safety management system, there remains no generally agreed upon means by which this should be accomplished. In fact, few tools exist that allow an operator to make such measurement with any substantial practical value. Data envelopment analysis (DEA) has been used extensively to compare organizational production efficiency, with techniques such as stochastic frontier analysis or the double-bootstrapped form of DEA infusing the process with a stochastic element. This study applies native DEA methodology toward the evaluation of safety culture and safety program effectiveness among similar organizations and proposes future research directions utilizing stochastic DEA methods toward achieving more reliable, generalizable results upon which statistical inferences may be based.

This research produces and tests a model for the evaluation of efficiency in creating a positive safety culture along an estimated efficiency frontier using DEA and further extends the model for future research in applying double-bootstrapped DEA methods to the problem. The results of this study offer evidence that DEA is a viable means of comparison of safety culture measures among organizations or organizational units, and show that it provides a tool for empirically-based prioritization of safety climate intervention at an operational level.