A number of logistical issues complicate the conduct of child disaster research. Like studies of adults, much of the child research has used a single cross-sectional assessment of non-representative samples, it fails to consider pre-disaster contribution to post-disaster problems, and it leaps to unwarranted causal conclusions from results providing mere associations. Despite concern about the accuracy of parental report and concern about children's understanding of certain terms, most child studies have used a single source of information - either the children themselves or their parents. As the field matures, greater attention to the sophistication of research methods and design will increase our understanding of children in the context of disasters.
Pfefferbaum, Betty, and Carol S. North. 2008. "Research with Children Exposed to Disasters." International Journal of Methods in Psychiatric Research (December): 49-56. http://www.ncbi.nlm.nih.gov/pubmed/19035441