There are limitations to our findings. While our analyses included adolescents 15–21 and young adults 24–36, they did not include individuals age 22–23. However, given the similarity of findings across the cohorts, we might expect a similar pattern of results, and we saw no age differences across our latent class solutions. In addition, seeking smoking cessation treatment was highly associated with nicotine withdrawal severity in both the adults and in the adolescent boys, but that information was not gathered from the girls at the time of their initial interview. In future work, we can examine this association with information provided through follow-up assessments of this cohort. Our assessments of nicotine withdrawal were retrospective in nature. While the degree to which recall of nicotine withdrawal symptoms reflect symptoms of nicotine withdrawal experiences in real-time is an on-going areas of study, previous psychometric studies of self-reported symptoms found high test-retest (Hughes et al., 1984; Tate et al., 1993) and moderate inter-rater reliability of nicotine withdrawal symptoms (Hughes et al., 1986). Retrospective reports of lifetime smoking history, more generally, show high agreement with real-time reports with agreements rates up to .74 after 14 years (Kenkel et al., 2003).
