The purpose of this thesis was to find out if existing information models could be used to describe information seeking in a disaster, what type of information seeking patterns exist in a disaster, and whether a disaster information...
moreThe purpose of this thesis was to find out if existing information models could be used to describe information seeking in a disaster, what type of information seeking patterns exist in a disaster, and whether a disaster information seeking model might be able to guide disaster communication practice. The research reviewed literature on human behaviour in disaster and information seeking in a disaster. While no models were found that illustrated the disaster information seeking process, a model developed by Mileti and Sorensen (1990) attempted to explain influences on risk communication. In addition, a model of problem-specific every day life information seeking (Savolainen 2008b) was reviewed that could possibly be adapted to disaster. This study uses these two models to develop a model of disaster information seeking.
The disaster information seeking model was improved using three research methods. They were semi-structured interviews of 51 disaster-experienced respondents, an online and mailed survey that attracted 345 respondents, and a focus group of six disaster communication practitioners and researchers. Each of these methods contributed to further development of the model, and the first two methods established a number of information seeking pathways possible in disaster. A feature of the interviews and survey was the influence of disaster type on information seeking behaviour – in particular, bushfire, cyclone, slow flood and flash flood. The focus group then investigated the usefulness of the model. Changes were made that the practitioners and researchers felt would make the model more useful and which were also supported by the literature and results of the interviews and survey.
Key findings from the interviews were specific to disaster type, which was expected after the literature review. Example information seeking pathways were established, with a dominant pathway for each disaster: flash flood was hearing about it from other people and then using television as a confirmation source; for slow flood, environmental cues alerted people to the possibility of a coming flood, followed by confirmation with other people; in a bushfire, other people were the alert source and the confirmation source; in a cyclone the alert sources were the Bureau of Meteorology (BOM) website and commercial radio, and the confirmation source was weather websites, including BOM. In flash flood, the key information need was to make sense of what had happened and how it had affected their own local community; in a slow flood, it was about the flood peak and how it would affect their own and family’s property; in a bushfire, it was about the location and path of the bushfire and the welfare of family and friends; in a cyclone it was the path of the cyclone before it hit and the level of damage after, and how friends and family had fared. The model developed from the literature review was subject to some change after the analysis of the interviews.
The survey confirmed the disaster specific nature of information seeking. In a flash flood, other people and environmental cues were most prevalent alert sources, with the confirmation source news and weather websites. In a slow flood, television and radio were the key alert sources, followed by news and weather websites. People in a cyclone learned about it form a news or weather website and confirmed using a different source with this same form cluster; in bushfire the alert sources was environmental cues followed by an agency website; in a storm, other people then television or radio; earthquake was environmental cues and then other people, radio or emergency agency staff; tornado was television or news or weather website confirmed by other people or environmental cues; tsunami was other people directly then news or weather website; and mudslide was other people directly then news or weather website. There seemed to be differences between participants in information seeking relating to age, gender, household size, education, proximity to the disaster, and the type of community people lived in. Changes were made to the model as a result of the survey analysis.
The industry focus group confirmed the validity of the model, with ‘filters’ adopted in place of ‘effect of situational factors’, a personalisation trigger was added and the point at which people exit the information seeking process was also added, with an option to re-enter at any time.
As a result, a final model of disaster information seeking was established. It describes the process of information seeking, influences on choices of source and form, and plots the triggers for either action or return to further information seeking. It also accounts for the importance of information sources and the tendency for people to return to a few trusted sources throughout the information seeking process. The model appears to provide a solid foundation for practice, and is ready for testing in further research.