z=b1b2/b12σb12+b22σb22, where b1 is the regression coefficient between the predictor and the mediator, b2 is the regression coefficient between the mediator and dependent variables, and σ2 is the square of the estimate of the standard error of the corresponding regression coefficient. Next, the common liability model was tested by examining whether age of alcohol use onset mediates the relationship between transmissible risk and CUD, and whether age of onset of cannabis use mediates the relationship between transmissible risk and AUD. A one-way ANOVA was then conducted to compare the magnitude of transmissible risk prior to substance use onset in youths who subsequently initiated alcohol use before and after first cannabis use, followed by Cox proportional hazard regression analysis to delineate the correspondence between TLI severity and age at the time of first alcohol and cannabis use. Furthermore, IRT methods were applied to the Drug Use Chart to delineate the probability of alcohol and cannabis use in relation to severity of overall substance use, taking into account 10 drug categories (alcohol, cannabis, crack/cocaine, opiates, amphetamines, sedatives, tobacco, hallucinogens, PCP, and inhalants;
