Use of wearable sensors to unveil roles of task demands-personal resources and burnout on performance of construction workers

This dissertation examines how task demands and personal resources affect construction workers’ productivity and safety performance. Additionally, the paper investigates the existence of a mediating mechanism of burnout in the relationship between job characteristics and workers’ performance. Through a quasi-experimental study, a modified “Job demands-resources model (JD-R)” for construction workers is validated at the task and individual levels. The modified model extends the discussion from job characteristics to performance consequences. In accordance with the Oldenburg Burnout Inventory, the model was designed to measure burnout by considering the two dimensions of exhaustion and disengagement in the context of construction workers from the non-service work category. The 22 subjects in the quasi-experiment comprised trainees in a pre-apprenticeship program and university students. They participated in multiple experiments that were designed to expose subjects to different levels of task demands and record different levels of personal resources; the final dataset for data analysis included 80 observations. Owing to the limited sample size and the explanatory nature of this dissertation research, the proposed model consists of indirect paths from task demands and personal resources to performance outcomes, and hypothesis testing is performed by applying partial least squares structural equation modeling (PLS-SEM). The mediation effect of exhaustion and disengagement was analyzed after including the direct path in the revised model, and utilizing only the significant paths among the indirect paths from task demands and personal resources to productivity and safety performance outcomes. The results indicate that exhaustion and disengagement have different relationships with performance outcomes. High burnout (exhaustion) and low disengagement (high engagement) subjects showed high productivity levels but low safety performance. Thus, there is a greater possibility that exhausted workers, who are depleted of mental and physical energy, will fail to comply with ergonomic safety, either intentionally or unintentionally. The combined model mixed survey and survey measurements results in a better overall predictive performance for the exogenous constructs than the model that only used either survey measurements or sensor measurements. Thus, the findings suggest that a human factors measurement method cannot replace another. Application of both survey and sensor measurements to human factor variables in the JD-R, burnout, and performance models is necessary for scientific construction workforce management in the construction industry. The dissertation contributes to the research stream in various ways. The key contribution of this dissertation research is the use of a scientific approach to evaluate construction workers’ physical strain and psychological stress. Further, it assesses the effects of such phenomena on their task- and individual-level performance. The study includes and measures both productivity and safety performances to provide insights to improve them and unveil the interrelated mechanism between productivity and safety, which have been discussed separately in prior studies. Specifically, the study measures the unique characteristics of construction performance on the two dimensions of safety (specifically ergonomic safety) and productivity and examines the daily (acute) burnout at the individual and task levels. Second, the study illustrates how the individual-level JD-R and burnout research can be conducted in the construction industry. The JD-R has been applied to a variety of industries, but not in contexts covering both safety and productivity in the construction industry. Indeed, this study increased the understanding of dynamics between safety and productivity. Third, the study extensively uses and uniquely describes the methodology applications of wearable sensors, including physiological status monitors and activity/sleep trackers for construction workforce management research and the application of PLS-SEM, an emerging tool in management research. By doing this, this dissertation provides detailed theoretical and managerial implications and also a discussion on the scope of future studies.