Results for 160 samples of disaster victims were coded as to sample type, disaster type, disaster location, outcomes and risk factors observed, and overall severity of impairment. In order of frequency, outcomes included specific psychological problems, nonspecific distress, health problems, chronic problems in living, resource loss, and problems specific to youth. Regression analyses showed that samples were more likely to be impaired if they were composed of youth rather than adults, were from developing rather than developed countries, or experienced mass violence (e.g., terrorism, shooting sprees) rather than natural or technological disasters. Most samples of rescue and recovery workers showed remarkable resilience. Within adult samples, more severe exposure, female gender, middle age, ethnic minority status, secondary stressors, prior psychiatric problems, and weak or deteriorating psychosocial resources most consistently increased the likelihood of adverse outcomes. Among youth, family factors were primary. Implications of the research for clinical practice and community intervention are discussed in a companion article (Norris, Friedman, and Watson, this volume).

The authors evaluated the impact of receiving social support on subsequent levels of perceived social support and psychological distress in 2 independent samples of victims of severe natural disasters: Hurricane Hugo (n = 498) and Hurricane Andrew (n = 404). A social support deterioration deterrence model was proposed that stipulated that postdisaster mobilization of received support counteracts the deterioration in expectations of support often experienced by victims of major life events. LISREL analyses of data collected 12 and 24 months after Hugo and 6 and 28 months after Andrew provided strong evidence for the hypothesized model: Perceived support mediated the long-term effects on distress of both scope of disaster exposure and postdisaster received support. Theoretical and application issues of social support are discussed.

SummaryDisasters typically strike quickly and cause great harm. Unfortunately, because of the spontaneous and chaotic nature of disasters, the psychological consequences have proved exceedingly difficult to assess. Published reports have often overestimated a disaster's psychological cost to survivors, suggesting, for example, that many if not most survivors will develop posttraumatic stress disorder (PTSD); at the same time, these reports have underestimated the scope of the disaster's broader impact in other domains. We argue that such ambiguities can be attributed to methodological limitations. When we focus on only the most scientifically sound research-studies that use prospective designs or include multivariate analyses of predictor and outcome measures-relatively clear conclusions about the psychological parameters of disasters emerge. We summarize the major aspects of these conclusions in five key points and close with a brief review of possible implications these points suggest for disaster intervention.1. Disasters cause serious psychological harm in a minority of exposed individuals. People exposed to disaster show myriad psychological problems, including PTSD, grief, depression, anxiety, stress-related health costs, substance abuse, and suicidal ideation. However, severe levels of these problems are typically observed only in a relatively small minority of exposed individuals. In adults, the proportion rarely exceeds 30% of most samples, and in the vast majority of methodologically sound studies, the level is usually considerably lower. Among youth, elevated symptoms are common in the first few months following a high-impact disaster, but again, chronic symptom elevations rarely exceed 30% of the youth sampled.2. Disasters produce multiple patterns of outcome, including psychological resilience. In addition to chronic dysfunction, other patterns of disaster outcome are typically observed. Some survivors recover their psychological equilibrium within a period ranging from several months to 1 or 2 years. A sizeable proportion, often more than half of those exposed, experience only transient distress and maintain a stable trajectory of healthy functioning or resilience. Resilient outcomes have been evidenced across different methodologies, including recent studies that identified patterns of outcome using relatively sophisticated data analytic approaches, such as latent growth mixture modeling.3. Disaster outcome depends on a combination of risk and resilience factors. As is true for most highly aversive events, individual differences in disaster outcomes are informed by a number of unique risk and resilience factors, including variables related to the context in which the disaster occurs, variables related to proximal exposure during the disaster, and variables related to distal exposure in the disaster's aftermath. Multivariate studies indicate that there is no one single dominant predictor of disaster outcomes. Rather, as with traumatic life events more generally, most predictor variables exert small...

The authors examined social causation and social selection explanations for the association between perceptions of social support and psychological distress. Data came from a sample of 557 victims of natural disaster in Mexico. Structural equation modeling analyses indicated that social causation (more social support leading to less posttraumatic stress disorder [PTSD]) explained the support-to-distress relationship in the earlier postdisaster phase, 6 to 12 months after the impact. Both causal mechanisms emerged as significant paths in the midpoint of the study (12 and 18 months). Only social selection (more PTSD leading to less social support) accounted for the support-to-distress relationship at 18 to 24 months after the event. Interpersonal and social dynamics of disasters may explain why these two contrasting causal mechanisms emerged over time.