Factor structure of the SDS has also been examined in samples which included participants with drug dependence, psychosis and long-term cannabis dependence [19,20,26]. Each showed that a single factor accounted for a considerable portion of the total variance (48.4 – 56.8%). Consistently, all five items of the SDS had high positive correlations on the factor score of greater than .50. Convergent validity of the SDS was demonstrated by reporting moderate to high significant correlations between the scale's total score and the estimated quantity of cannabis use, frequency of use and the number of DSM-IV dependence criteria [19,21]. In the study of Swift, Hall et al. [26] significant correlations between SDS and quantity and frequency of use as well as age at first use and total duration of use were not found. However, in their sample of long-term cannabis users they could show a substantial agreement between the SDS score and the respondents' belief that their cannabis use caused problems (kappa = .44).
