Demographic information and pre-test comparisons are presented in Table 1. The first set of analyses tested for differences among CNR1 (T/T versus T/C) and FAAH (C/C versus C/A) genotype groups on baseline demographics and marijuana use variables that might confound the main analyses. There were no differences for either of these genes on age, race or gender. There were, however, differences on a few marijuana use measures, such that CNR1 T/C individuals scored significantly higher on the MDC and reported significantly higher quantity and frequency of marijuana use per month and day, respectively. These variables were not used as covariates because this study utilized an intermediate phenotype approach (aka, endophenotype), and as such it is expected that the genetic variable(s), the intermediate phenotype(s) and the clinical phenotype(s) are all related. The premise of this approach is that genetic variables contribute to withdrawal and cue-elicited craving, which contributes in turn to frequency of use and dependence. The correlation between MDC and the withdrawal score after abstinence was r(100) = 0.27, P < 0.05, and the correlation between MDC and craving after cue exposure was r(102) = 0.19, P = 0.058.