OPENING PLENARY: "DISASTER RESEARCH: PAST INNOVATIONS AND FUTURE DIRECTIONS."

Regardless of whether they are natural in origin or human-caused, disasters are extremely complex events. They affect large numbers of people simultaneously. They generate an array of stressors of varying degrees of intensity that interact with multiple pre- and post-event characteristics of the person and environment to produce diverse outcomes that evolve over time. Most disaster studies examine the effects of a particular event that occurred at a particular time to a particular population in a particular place. These elements of time, population, and place may serve to organize the methodological challenges, regardless of whether the research aims to describe or explain the disaster's effects or aims to study approaches that alleviate those effects.

With regard to time, motion pictures, rather than snapshots, are needed to portray the aftermath of disasters. Prospective, longitudinal designs are best, and advancements such as growth curve modeling have much to contribute to this field. The suddenness and unpredictability of disasters makes it difficult for researchers to enter the field quickly, and many systemic barriers to rapid response research exist. Intervention research must control for the natural course of recovery and deal with changing needs over time. Because of the sense of urgency that accompanies disasters, concerns about experimental designs and scientific rigor often take a back seat to provider beliefs, consumer demands, and clinical necessities.

Because disasters happen to entire communities, the ability to describe the afflicted population is the sine qua non of disaster research. Whether or not comparison groups are necessary and how to go about selecting them are long-standing methodological issues. Innovations that improve sample representativeness by reducing, or modeling, non-response, attrition, and missing data are critical. There is also the issue of how well the entire body of disaster research represents the entire population of disaster victims. Children are understudied because they are difficult to sample, recruit, and assess. Most disasters occur in developing countries, but most research occurs in developed countries and, within developed countries, minority and marginalized communities are also understudied. Methods for improving the validity of assessment instruments for different age ranges and cultural groups are important for disaster research, as they are for other areas of trauma research. Also important are methods for analyzing non-independent observations. For the most part, our research strategies have not acknowledged the connectedness of disaster victims, and we study individuals as if they each experienced separate stressful life events. Hierarchical linear modeling has much to offer the disaster field; indeed the entire field of ecological assessment holds promise for elucidating the transactions of individual, family, and community recovery. Population issues are equally important for intervention research. Studies of post-disaster interventions must consider developmental and cultural issues from the outset and be designed to have adequate power to examine effects for specific age or cultural groups. Methodologies that address effectiveness may be more important than methods that determine efficacy in this context. Solutions to the issue of self-selection biases have been demonstrated in a number of methodological articles. The development of interventions that target community as well as individual recovery is limited by the methodological challenges in assessing community-level change.

Last, but not least, methodological innovations are needed that help disaster researchers do a better job of incorporating aspects of place into their work. Human behavior must be understood in context. Again, advances in ecological assessment may be helpful for this purpose, and there are excellent examples from others fields, such as criminal justice. Qualitative approaches provide an alternative, and much more needs to be done with regard to integrating qualitative and quantitative approaches in disaster research.