Understanding Disaster Warning Responses.

When threatened with some type of disaster, how do people respond? What are the social factors that constrain their responses? Receiver characteristics, message characteristics, and social contexts are explained and related to variations in disaster warning responses. Finally, two components of a vision for the future are described: (1) disaster event taxonomies, and (2) implemented social policies.

On Saturday, Sept. 8, 1900, a massive hurricane struck opulent, but vulnerable, Galveston Island which sits in the Gulf of Mexico off the mainland from Houston, Texas. Warnings were near nonexistent and at least 6000 people perished (McComb, 1986). During the decades since this tragedy, U.S. coastal areas have become even more vulnerable due to increased population densities and explosive development (Mileti, 1999). These same social forces also place more persons at risk because of other hazards like floods, volcanoes, tornadoes, earthquakes, and the like (FEMA, 1994). Yet, fewer lives are lost each decade largely because of the implementation of improved warning systems. This article presents a brief summary of what has been learned about disaster warning responses and what yet remains unknown through the examination of four topics: (1) threat denial; (2) warnings as social processes; (3) networks of social constraints; and (4) a vision for the future.

THREAT DENIAL

The first principle in understanding disaster warning responses is to recognize explicitly that the initial response to any warning is denial. Recent studies have underscored the complexity of such experiences (e.g., Mileti and Fitzpatrick, 1993). For example, Rogers (1997) compared resident risk perceptions following a chemical plant fire in Odessa, Texas, and a 7-year controversy over the siting of a hazardous waste incinerator in LaPorte, Texas. By comparing survey results obtained in 1993 with those from a year earlier from the same individuals, Rogers documented the impact of these two contrasting types of events. About seven out of 10 (67% in Odessa and 70% in LaPorte) "indicated that their estimates of the likelihood of risk had changed during the study period" (Rogers, 1997, p. 70). But despite this claim, comparisons of the risk estimates people offered i.e., the chances of a significant release of potentially toxic materials, at the two points in time actually varied little. This was true even for t hose in Odessa who were evacuated at the time of the chemical plant accident. Conclusion? Rogers summarized it this way. "These results also indicate that there is a strong component of inertia holding perceived risk at stable preconceived levels" (Rogers, 1997, p. 76).

Despite these and related findings, limited evidence indicates that consistent and repeated experiences with hurricanes has elevated both understanding and more realistic perceptions of risk (Cross, 1990). But what about areas where most people live? In most U.S. communities, disaster events reflecting any specific type of hazard are relatively infrequent. In such places, the extent of threat denial is shocking.

The statement brings to mind field interviews (n = 57) I conducted recently in Sterling, Colorado, a small (1996 population: 11,268) agriculturally based community about 130 miles northeast of the Denver metropolitan area. How intense is threat denial to disaster warnings? Consider the following comments I reconstructed from my field notes with a local businessman whose business had been operating continuously for over four decades.

By way of context, the flood event being discussed resulted from intense overnight rainfall in the hills to the northwest of this community. The rain started during the night on Tuesday, July 29, 1997. Water moved southward during the early morning hours and flooded most homes in the tiny town of Atwood (population 1996 = 180) about 7:00 a.m. Atwood sits just northeast of where Pawnee Creek flows under U.S. Highway 6. From …