of electronic communications will create a crisis of overload for the telephone system.

• Every attack—whether successful or thwarted—can be expected to enhance efforts to prevent further attacks of the same kind. A simple but telling example is the instituting of random shoe inspections of airline passengers after the aborted shoe-bomber incident in December 2001.

• If the attack is believed to have been avoidable, and the agents responsible for its avoidability are identified or suspected, a season of scapegoating, public investigations of culpability, and calls for punishment will ensue.

• If agencies of public order (police, National Guard, military) and rescue agencies (firefighters, Red Cross, volunteer workers) are perceived to have been ineffective or improperly coordinated, scapegoating will be directed toward these agencies as well.

• Contrariwise, there will be an identification and adoration of heroes in crises. This effect will also decrease if attacks become repetitive.

Most disasters are both sudden and ephemeral, and immediate responses quickly give way to a wide variety of long-term recovery and rebuilding activities. Therefore research on immediate disaster responses generally relies on hastily assembled journalistic reports and after-the-fact accounts based on participants’ recollections. Both types of sources are subject to selectivity and distortion. Teams of behavioral and social science researchers collecting data on the spot and analyzing it in the context of established knowledge about disaster situations would supplement and likely improve on existing ways of generating information about disaster response. Some universities have a tradition of such fire-brigade research, but efforts should be made to expand and systematize it.

Recommendation 9.6 (Research): Relevant research agencies (universities, think tanks, or government) should establish the capacity to move quickly to the scene of a disaster and study immediate responses while they are occurring.

Analysis of preparedness, warning, and response tends to rest on the assumption of an undifferentiated community or public. Research on disasters, however, has revealed that individuals and groups differ both in readiness and response according to previous disaster experience, ethnic and minority status, knowledge of the local language, level of education, level of economic resources, and gender (Tierney, Lindell, and Perry, 2001). Research on these and other differences should be extended and deepened, and it should be taken into account when designing systems of preparedness, warning, and response to terrorist attacks and other disaster situations.

Recommendation 9.7 (Research): Research on how different individuals and groups prepare for and respond to crises should be extended and deepened.