Analyses indicated the best fitting solution for women and men were 3-class solutions, while the best fitting solutions for girls and boys were 2-class solutions. Models could be constrained across adult women and men (X226=38.0; p >.05; figure 1. a.), but could not be constrained across girls (figure 1. b.) and boys (figure 1. c.; X217=38.6; p <.05). Girls in the mild class reported significantly higher rates of withdrawal-related depressed mood and increased appetite compared to boys in the mild class (4.2% vs. 0% and 45.0% vs. 18.3%, respectively). The pattern that emerged across all the groups was consistent with a severity continuum for nicotine withdrawal, where the probability of an individual endorsing each nicotine withdrawal symptom was significantly higher with each respective class. Evidence for this included, with a single exception, the lack of overlap of standard error bars (i.e., no lines that overlapped or crossover). The only borderline exception to this case was in the adults, where the endorsement probability for withdrawal-related increased appetite for the moderate class overlapped that for the severe class. This moderate class in
