PSYCHOSOCIAL IMPACT OF NATURAL DISASTER: AN ARCHIVAL STUDY

The controversy surrounding the existence of adverse psychosocial effects of natural disasters was reviewed as an inevitable artifact of the confluence of investigations from the psychological and sociological research traditions. Together, the findings of these research traditions suggest a demand adaptation model of disaster behavior that emphasizes both social and psychological determinant processes. The presuppositions of this model were articulated in the methods of the present study, which was designed to describe the longitudinal effects of natural disaster on a sample of archival measures. An interrupted time series design was used to assess the longitudinal effects of two disaster events in two comparable New York State counties. The archivally collected dependent variables were drawn from three domains including psychiatric morbidity, service utilization, and social adjustment. Based on the postulates of a demand adaptation model of disaster behavior it was hypothesized that the stresses of post-disaster adjustments would be most marked in the relatively subtle measures of social disruption and service utilization, while measures of psychiatric morbidity were expected to demonstrate little change. The data series were analyzed to assess changes from baseline trends in the period following disaster impact. The analyses also included identification of trends and autocorrelation within the data, since these factors are typically overlooked and tend to confound longitudinal disaster studies. Despite numerous liabilities in the data base and compromises in design rigor, the results of this study demonstrated that major natural disasters can result not only in increased social disruption but in adverse mental health effects as well. The findings in the domain of service utilization were equivocal however, and suggested that the measures chosen here may not have been appropriate for the purposes of this study. Analyses of the statistical properties of the data revealed the existence of both non-stationarity (trend) and autocorrelation in most of the data series sampled; two findings which cast suspicion upon archival studies that do not control for these phenomena. Most importantly, the results demonstrated the necessity of considering the unique qualities of the disaster agent, the idiosyncratic features of the disaster community, and the specific measures selected for study when interpreting archivally collected data in disaster research