TRACEr-RAV: modification of the 'technique for retrospective analysis of cognitive errors' for Australian rail use

There is currently no widely accepted human error identification (HEI) tool used in Australia to identify and classify errors associated with rail accidents and incidents. The aim of this paper is to outline an attempt to develop a HEI tool for Australian use. The Technique for the Retrospective and Predictive Analysis of Cognitive Errors (TRACEr) is a HEI tool initially developed for air traffic control and subsequently revised and adopted in the UK as a rail-specific tool for train driving (TRACEr-Rail). In applying TRACEr-rail to an Australian sample of incidents and accidents, only moderate inter-rater reliability was demonstrated. The tool was also unable to classify adequately all the errors (and factors associated with errors) extracted from incident reports. In an attempt to improve the consistency with which the tool is applied and the tool's capacity to identify and categorise train driver errors in Australia, the tool was modified to become TRACEr-RAV (TRACEr-for Rail, Australian Version). Changes to TRACEr-Rail included the addition and omission of some error categories, revision of many of the category definitions, the provision of more appropriate examples, and the adoption of more Australian terminology. This paper outlines the difficulties encountered while using TRACEr-Rail and the changes made to the tool. It also describes the process of design for a study aimed to establish the reliability and usability of our new tool, TRACEr-RAV.