Are Visual Search Patterns Predictive of Hazard Recognition Performance? Empirical Investigation Using Eye-Tracking Technology

Poor hazard recognition is a widespread issue in the construction industry. When construction hazards remain unrecognized, workers are more likely to indulge in unsafe behavior, experience unanticipated hazard exposure, and suffer catastrophic injuries. To improve our understanding of why construction hazards remain unrecognized, the current study examined hazard recognition as an everyday visual search task—similar to an individual searching for a product in a supermarket, a radiologist examining a radiograph for tissue abnormalities, or a security personnel screening baggage at an airport terminal. More specifically, the research used eye-tracking technology to examine the relationship between visual search patterns adopted by workers while participating in a hazard recognition activity and the resulting performance levels (i.e., hazard recognition performance). The research also focused on testing the effects of introducing a recently developed personalized training intervention on visual search patterns adopted by workers—and the subsequent hazard recognition performance. Visual search patterns examined in the study included search duration, fixation count, fixation spatial density, and others identified from past research. The results reveal that several quantifiable visual search patterns are predictive of superior hazard recognition performance. For example, workers who spent more time examining the workplace for safety hazards (i.e., search duration) recognized a larger proportion of hazards. Likewise, workers that devoted higher levels of attention through a larger number of fixations (i.e., fixation count) and longer fixation durations (i.e., fixation time), and those that distributed their visual attention more broadly across the work area (i.e., fixation spatial density), demonstrated superior performance. The findings also suggest that the adoption of the personalized intervention can lead to improvements in visual search patterns and hazard recognition performance among workers. The present research will be useful to diagnose and remedy search weaknesses demonstrated by workers (i.e., human factors) that are associated with poor hazard recognition levels.