Some other limitations of the present study are worth noting. First and foremost, the present sample was ascertained from three family studies of substance use disorders for the express purpose of identifying genetic variants for alcoholism, nicotine and cocaine dependence and related psychopathology. Hence, the psychometric analyses may not generalize to other cohorts with different ascertainment criteria. Second, while we were able to include a measure of cannabis withdrawal in the analysis, the symptoms and diagnostic scheme (i.e., 3 or more of 7 withdrawal symptoms) used to assess withdrawal do not conform to those in DSM-5. This was unavoidable since all studies predated the DSM-5 by a considerable number of years. However, analyses using the DSM-5 criteria in an independent twin sample do not indicate any evidence for genetic influences on DSM-5 withdrawal that do not overlap with DSM-IV cannabis abuse/dependence (Verweij et al., 2012).
