Previous studies of adult cigarette and marijuana use, adopting the CCC model approach, found a similar pattern of results to the present study [24,25]. Maes and colleagues [24] reported that the majority of the variance in quantity of cigarette use was explained by factors influencing initiation of cigarette use. However, unlike the present study, they reported that the beta coefficient could not be set to 1 without a significant drop in model fit, and concluded that two separate but related liabilities for initiation and progression characterized their data most effectively.
