Abstract
Studies of the accuracy of long-term recall for information about smoking status have tended to conclude that this information is consistent and reliable. Estimation of consumption levels have been found to be less reliable and can be influenced by current consumption levels. Results of this longitudinal study indicated that children's and adolescents' 2-year recall of smoking status was inconsistent and that children's reports of age of initiating smoking showed an unacceptably high level of misclassification, particularly if they had not smoked in the last 2 years. Children's and adolescents' 2-year recall for the age at which they first smoked were imprecise and suggested that those who had smoked in the last 2 years retained a perspective as they grew older of having started in the last few years. It is recommended that the period of recall for children and adolescents be restricted to no more than 1 year.