Testing the Reliability of Accident Analysis Methods: A Comparison of Accimap, Stamp-Cast and Accinet

Authors:
Preprint
Publisher website
Google Scholar
Abstract
Accident analysis methods are used to model and understand the multifactorial cause of adverse incidents. Methods such as AcciMap, STAMP-CAST and more recently AcciNet, are systemic approaches that support the identification of new safety interventions across multiple sociotechnical system levels. Despite the growing popularity of systems-based accident analysis methods, little is known about how reliable and valid they are when used to describe, model and classify contributory factors and relationships. This is the second of two studies in the Journal to examine and compare the reliability and validity of accident analysis methods. Here, we conducted an intra-rater (between analyst) and inter-rater (between analyst) reliability assessment of AcciMap, STAMP-CAST and AcciNet, and used the Signal Detection Theory paradigm to analyse the complete dataset. The findings indicate that STAMP-CAST produced a significantly higher inter-rater reliability coefficient relative to AcciMap and AcciNet. No statistically significant or practically meaningful differences were found between the three methods in the intra-rater reliability analyses. Implications and directions for future research are discussed.

This publication has 44 references indexed in Scilit: