Brian H.W. Guo

Developing Leading Indicators to Monitor the Safety Conditions of Construction Projects

AbstractThis paper presents a conceptual framework for developing leading safety indicators for the construction industry. The framework clarifies the nature of the indicators in terms of definition, purpose, and attributes. A pragmatic method for systematically identifying a set of leading indicators for construction projects is proposed. The method consists of four steps: conceptualization, operationalization, indicator generation, validation, and revision, and emphasizes two functions of leading indicators: informative and decision aiding. First, leading indicators must be able to provide information about the state of construction safety. Second, they must be able to help decision makers take remedial actions. These two functions should promote double-loop learning, reflecting any existing safety model and facilitating the construction of a new one through ongoing validation. A hypothetical example is provided to illustrate the entire development process. The conceptual framework, together with the de...

Knowledge, Attitude, and Practice of Design for Safety: Multiple Stakeholders in the Singapore Construction Industry

AbstractThis paper aims to investigate the design for safety (DfS) knowledge, attitude, and practice (KAP) of multiple stakeholders, including architects, civil and structural (CS) engineers, mecha...

A System Dynamics View of Safety Management in Small Construction Companies

Due to unique characteristics of small construction companies, safety management is comprised of complex problems. This paper aims to better understand the complexity and dynamics of safety management in small construction companies. A system dynamics (SD) model was built in order to capture the causal interdependencies between factors at different system levels (regulation, organization, technical and individual) and their effects on safety outcomes. Various tests were conducted to build confidence in the model`s usefulness to understand safety problems facing small companies from a system dynamics view. A number of policies were analyzed by changing the value of parameters. The value of a system dynamics approach to safety management in small construction companies is its ability to address joint effects of multiple safety risk factors on safety performance with a systems thinking perspective. By taking into account feedback loops and non-linear relationships, such a system dynamics model provides insights into the complex causes of relatively poor safety performance of small construction companies and improvement strategies.

Factors influencing unsafe behaviors: A supervised learning approach

Despite its potential, the use of machine learning in safety studies had been limited. Considering machine learnings advantage in predictive accuracy, this study used a supervised learning approach to evaluate the relative importance of different cognitive factors within the Theory of Reasoned Action (TRA) in influencing safety behavior. Data were collected from 80 workers in a tunnel construction project using a TRA-based questionnaire. At the same time, behavior-based safety (BBS) observation data, % unsafe behavior, was collected. Subsequently, with the TRA cognitive factors as the input attributes, six widely-used machine learning algorithms and logistic regression were used to develop models to predict % unsafe behavior. The receiver operating characteristic (ROC) curves show that decision tree provides the best prediction. It was found that intention and social norms have the biggest influence on whether a worker was observed to work safely or not. Thus, managers aiming to improve safety behaviors need to pay specific attention to social norms in the worksite. The study also showed that a TRA survey can be used to extend a BBS to facilitate more effective interventions. Lastly, the study showed that machine learning algorithms provide an alternative approach for analyzing the relationship between the cognitive factors and behavioral data.