Challenge to Collect Empirical Data for Human Reliability Analysis—Illustrated by the Difficulties in Collecting Empirical Data on the Performance-Shaping Factor Complexity

Abstract

Abstract This paper discusses the challenges with collecting positivistic empirical data (objective, observable, reliable, replicable, experimental, and true) in human reliability analysis (HRA), and illustrates it by presenting the difficulties in collecting empirical data on the performance-shaping factor (PSF) complexity. The PSF complexity was chosen to illustrate the difficulties with empirically collecting data because it is included in many HRA guidelines and it has been discussed as an important PSF to understand error rates in large accident scenarios. This paper discusses the challenges with collecting empirical data from a pure positivistic paradigm with experiments, as well as from literature reviews and data from event reports, training, and operations. The paper concludes that because of all the challenges with the positivistic empirical data collections methods in HRA, we should discuss whether experts' judgements could be a better approach to obtain HRA data and error rates. In a postpositivistic view, qualitative data or experts' judgments could also be looked at as empirical data if the data were collected in a systematic and transparent way.