Self-reported drinking behavior is a potential limitation of the study. This was a sample from a longitudinal survey of media exposure and adolescent behavior. In addition, this sample, while national, might not be representative of groups with higher attrition in the survey over time, such as poorer or minority adolescents. Also, because the study was cross-sectional, no causality can be implied, Dr. Tanski said.
Dr. Tanski said that she had no relevant disclosures.