Citation

Niu Y, Fan Y, Gao Y, Li Y. Safety Sci. 2023; 161: e106082.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.ssci.2023.106082

PMID

unavailable

Abstract

Although the research into causal effects has always been the core part of safety science, the formal mathematical method to infer causal effects is quite new. With the rapid development of big data technology, safety research based on experience changes to data-driven research. This change improves the reliability and accuracy of safety decision-making to a certain extent. However, researchers pay too much attention to the amount of data and computing power, accelerating the separation of prior knowledge and data. We proposed a systematic method suitable for safety science to carry out quantitative causal inference to study causal problems such as the effectiveness of safety countermeasures. This method aims to integrate safety-related prior knowledge and data organically and provide essential guidance for research design and data collection. Finally, an empirical study was conducted to illustrate how the approach can accurately estimate the rigorous causal relationship between factors and how it can help researchers understand the complex causal mechanism behind the safety or accident phenomenon.

Language: en

Keywords

Causal effects; Causal inference; Prior knowledge; Safety decision-making