Three approaches to adolescent smoking detection: a comparison of "expert" assessment, anonymous self-report, and co measurement.

A new method designed to assess adolescent smoking was recently developed and applied by researchers in the former Soviet Union. As originally applied in Lithuania and Russia, two official student leaders from each class were chosen to estimate smoking prevalence rates for their class, and so the method became labeled as the "expert assessment." Their findings suggested that this "expert assessment" method agreed quite well with school-wide prevalence estimates based on a survey of the entire school population. The purpose of the present study was to determine whether student representatives, selected at random and coached to accurately summarize the smoking habits of their classmates, could provide accurate estimates of the smoking prevalence rate in their school. To accomplish this purpose, estimates from the student representatives were compared to those from an anonymous survey and from measurement of carbon monoxide in expired air. The results indicated that the self-report and carbon monoxide data were in close agreement, while the student representatives substantially overestimated the prevalence of smoking in their school, in spite of the coaching provided according to the "expert assessment" protocol. Additional research on alternative methods to select student representatives, such as those based on a sociometric analysis, may yet result in a low-cost method to assess schoolwide levels of smoking in adolescent populations.