Aggregating data from multiple informants in child psychiatry epidemiological research.

One of the issues confronting the field of child psychiatric epidemiology is devising ways to aggregate discrepant reports from multiple informants obtained in structured diagnostic assessments of children and adolescents. The present report uses data from an epidemiological community survey to attempt to address this issue. The results obtained through statistical procedures designed to identify the "optimal" informant for specific symptomatic criteria and diagnoses do not seem to offer a clear-cut advantage over a simple combinatorial rule that identifies symptomatic criteria as positive when they are acknowledged as positive by either parent or child informants.