Age of smoking milestones: longitudinal inconsistencies and recanting.

PURPOSE: To determine (1) how reports of the ages of first cigarette smoked and daily smoking onset change from adolescence through emerging adulthood and into young adulthood and (2) what predicts reporting inconsistencies and recanting for both smoking milestones.
METHODS: Multinomial logistic regression models compared relative risks of the following: (1) consistent reporting of milestone age (reference group); (2) recanting at either subsequent wave; or (3) inconsistent reporting of age in at least one subsequent wave, using data from Waves I, III, and IV of the National Longitudinal Study of Adolescent to Adult Health.
RESULTS: Instability and forward telescoping between adolescence and emerging adulthood leveled off by young adulthood. For smoking first cigarette, those who started younger had more inconsistencies and recanting than those who started later, as did African-American, Latino, and Asian respondents compared with non-Latino white respondents. Native American respondents also had higher relative risks of recanting, as did those with low parental education. Males were more inconsistent than females. Depression, same-sex attractions or relationships, and family structure were not associated with reporting stability. Binge drinking, marijuana, and other illegal drugs were associated with lower levels of recanting. For age of daily smoking, starting older versus younger, sex, race, ethnicity, and use of marijuana were significant predictors of report stability.
CONCLUSIONS: Stage of life may influence forward telescoping in smoking self-reports. Stability of reports of adolescent smoking by emerging and young adults in the United States appears biased by age of onset, sex, race, and other substance use.