on “the first 5 times you ever drank,” and items contributing to TOLMX concern the “period when you drank the most.” The score of each measure was computed by summing the number of drinks required to produce an effect and dividing by the number of effects endorsed. The SRE has been shown to have good internal consistency and test–retest reliability to successfully identify people who had low response to alcohol in a laboratory challenge test and to be associated with AD diagnosis in several populations. Because of non-normal distributions of the regression residuals, we log transformed values for ISENS and TOLMX. MAX24 refers to the largest number of drinks an individual reported ever having consumed in 24 hours. The withdrawal severity factor score was based on ten symptoms in the SSAGA interview (such as hands trembling, feeling anxious following cessation or reduction of drinking). To account for the possible non-equal contribution of each symptom to withdrawal severity, a factor analysis was conducted (for details see Kuo et al., 2006). A factor score of withdrawal severity for each individual was derived based on the item loadings on one major factor, which accounted for 70% of the variance in these symptoms based
