Enhancing construction hazard recognition and communication with energy-based cognitive mnemonics and safety meeting maturity model: Multiple baseline study

Nearly every safety management activity is designed around the fundamental and implicit assumption that the workforce is capable of identifying hazards before exposure. Unfortunately, research shows preliminary evidence that construction crews perform relatively poorly with respect to hazard recognition. This may be attributable, in part, to the diverse and dynamic nature of construction work. Consequently, many hazards remain unidentified, uncontrolled, and unmanaged until they are encountered by workers. To advance theory and practice in the area of construction hazard recognition and communication, a large-scale, multiphase research project was conducted with the following objectives: (1) devise a new hazard recognition method based upon the principles of cognitive mnemonics and that promotes hazard recognition skill; (2) build a maturity model to improve hazard recognition performance and encourage hazard communication; and (3) experimentally measure improvement in hazard recognition and communication resulting from the new program. The first two objectives were achieved by using the nominal group technique driven by input from an expert panel of 14 industry professionals from organizations with world-class safety records and an average of 26 years of safety management experience. A new method for measuring the quality of prejob safety meetings was devised in a series of eight face-to-face meetings and teleconferences among the experts over a 1-year period. The null hypothesis that the strategy does not improve the proportion of hazards identified and communicated before the start of construction was experimentally tested with six crews using the multiple baseline testing approach. This is a series of concurrent longitudinal A-B designs that are time-lagged among independent treatment groups. The obtained observations were analyzed using interrupted time series regression models. The results showed that workers were able to identify and communicate an average of 38% of hazards before the intervention was applied and that the new strategy caused a weighted overall level-change improvement in hazard recognition skill of 31% (p<0.01). This is the first known quasi-experimental test of a hazard recognition strategy in any industry and one of the first large-scale studies to employ multiple baseline testing with appropriate controls that optimize validity and statistical rigor.