A systematic review of eye-tracking in construction hazard recognition and risk assessment

Hazard recognition failures are among the most significant contributors to accidents in the construction industry, as workers may fail to visually inspect hazards, misinterpret observed cues, or underestimate risk under demanding task conditions. Eye-tracking technology has increasingly been applied in this area, as it can provide objective evidence of visual search behavior and risk perception. This study presents a systematic review of 46 peer-reviewed papers published between 2016 and 2025 on eye-tracking applications in construction hazard recognition and risk assessment. The review uses a three-stage approach that integrates bibliometric screening, scientometric mapping (using VOSviewer and CiteSpace), and qualitative thematic analysis. The results show that current research can be organized around four interrelated themes, including visual-attention mechanisms, individual differences, cognitive fatigue or mental-state effects, and the integration of immersive or multimodal physiological monitoring. The synthesis also reveals important methodological and practical limitations. For example, the evidence base remains strongly laboratory-centered, with 31 of 46 studies conducted in laboratory settings, only 3 studies conducted in field settings, and 24 studies relying on static images in assessment. These patterns indicate strong support for examining the visual-attention mechanism under controlled conditions, but limited evidence for transferability to active construction sites. Based on the synthesis, the study proposes the Attention-Moderators-Interventions-Outcomes (AMIO) framework and suggests that field validation, longitudinal retention assessment, standardized purpose-based metric selection, multimodal validation, implementation feasibility, cost-benefit evaluation, privacy-preserving sensing architectures, and worker-centered deployment are key priorities for future research.