Warning and Evacuation During a Mass Disaster: A Multivariate Decision Making Model *

Abstract

The development of an evacuation decision making model is explored here. Information obtained from a 20 percent areal sample of population who were in a disaster area in Israel were asked to reconstruct the events leading up to, during and after the crisis. Affective reactions and behavioral acts were recorded. Evacuation was not universal thereby providing the basis of a multivariate analysis whose aim was to generate a parsimonious logistic regression model so as to decipher what independent variables could best explain the decision to leave or stay. The parsimonious model, based on previous research, focused on major conceptual and empirical factors involved m an evacuation decision. The results indicated that a positive/negative decision to stay or leave a disaster area is dependent upon a specific aspect of the warning process, namely the means of information acquisition, its confirmation and the degree of support provided by neighbors and neighborhood social networks. These results point out the complexity of the model and stress how individual, family and community are all bound up in the evacuation process.