Development, use and usability of TRACEr-RAV (Technique for Retrospective Analysis of Cognitive Errors for Rail, Australian Version

The current study is part of a wider project examining the development of a version of TRACER-Rail (developed by the RSSB) for theAustralian rail context.As part of a wider project examining the reliability of error classification tools in rail, the current study examined the usability of the newly developedTRACER-RAV (Technique for the Retrospective Analysis of Cognitive Errors- for Rail, Australian Version). An earlier workshop study with railexperienced personnel found that despite changes to the tool that were intended to makeTRACER-RAV more user-friendly, and more comprehensive than TRACEr-Rail, it was rated as no more useable and proved to be no more reliable than the previous version of the tool. The lack of improvement in usability of TRACEr-RAV may have been observed because of the between subjects design adopted. Participants were given either TRACER-RAV or TRACER-Rail, and so were unable to reflect differences in usability of the tools because they only saw one version. Following further improvements to TRACER-RAV, the current study aimed to assess the usability of TRACER-RAV using a within subjects design. A sample of postgraduate students each completed two sessions of error classification - one with TRACER-Rail and the other with TRACER-RAV. Sessions were conducted in counterbalanced order. Following training with a tool, participants identified and classified errors from the executive summaries of 3 Australian rail incident reports. In the second session, they were trained to use the other tool, and classified errors from 3 new incident reports. Participants rated the usability of each tool at the end of the session by completing a questionnaire. Overall, TRACEr-RAV resulted in better usability ratings than TRACEr-rail, specifically in relation to the ease of understanding of instructions, ease of use of the recording form, and perceived independence of classification categories. TRACEr-RAV thus appears to be a more usable version of the tool, which should in turn reflect on the quality of classifications made.