Life Risk Analysis in Residential Building Fires

Abstract

This study focuses on the life risk investigations in residential building fires. A total of 419 cases in residential building fires during 1998-2000 in some area of Taiwan are examined. Contexts of investigations include possible influencing variables or factors related to the activities of pre-fire as well as during fire and the resulting life casualties. Possible influencing variables or factors are measured by two questionnaires that include both fire departments and the households who owned or lived in the fire-hit residential buildings. The resultant fire life risk (the dependent variable) in this study is a dichotomous variable with two categories, casualties and non-casualties. Factors such as fire-fighting situations, degree of difficulty of fire escape, and disabled people in the building are mostly significant to estimate the life risk. That disabled people exist in the building in case of a fire has the strongest effect on life risk. A statistical interpretation is given of the model-developed estimates in terms of the odds ratio. Given that particular identified characteristics considered in residential buildings exist or not, the changes of the probabilities of causing casualties in case of a fire are predicted. The findings show that logistic regression used in this research provides good meaningful interpretations that can be used for future safety improvements.