Abstract
When surveyed twice within a 6-month period, to what extent do young adolescents give inconsistent answers to questions about sexual behavior and substance use (cigarettes, alcohol, marijuana, inhalants, cocaine)? Data were collected from 1,575 urban African American and Hispanic students during fall and spring of 7th grade. For each behavior examined, less than 2% of the sample gave inconsistent answers within a survey and less than 7.5% did so over time. Retraction of baseline answers at follow-up was greater for rarer and more socially undesirable behaviors (e.g., cocaine versus cigarette use). Over-time inconsistencies were associated with lower reading comprehension and an overall reduction in risk behaviors at follow-up. [Translations are provided in the International Abstracts Section of this issue.]