Predicting the Human Response to An Emergency

Abstract

This paper presents a case study of an approach for predicting the human response to a domestic fire, using a combination of a talk-through technique (Kirwan and Ainsworth, 1992) and sequential analysis (Bakeman and Gottman, 1986). 20 participants were asked what actions they would take upon hearing a strange noise in their house, which they were later told was a fire. Each act was recorded and the results were compared to previous research in which people involved in real fires had been interviewed (Canter et al, 1980). A significant relationship was found between the frequency (Spearman's rho: 0.694, p<0.01) and sequence (Spearman's rho: 0.441, p<0.05) of acts in this study and those from the interviews with people involved in real fires. More work is needed to develop the approach, but this case study indicates that it might have use as a low-cost method which can be used to predict behaviour in an emergency.