Thus, for each study, data from each participant were combined into EEG scores in the alcohol and placebo/pre-drinking conditions, using the up to 3 variables, weights, and variances calculated in the divergence analysis in Study 1. Then for each study, participants’ post-drinking and placebo/pre-drinking scores were compared using a receiver-operator characteristic (ROC) analysis to assess the significance of recognition of alcohol [30]. Area under the ROC curve (AUC), sensitivity (correctly identifying the alcohol condition), and specificity (correctly identifying the placebo/pre-drinking condition) values were computed. The significance of the AUC was then determined by transforming the AUC into a z-score with respect to the null hypothesis of AUC = 0.5, and referencing the z-score to the normal distribution, i.e. z score = AUC minus 0.5 divided by the standard error for AUC = 0.5. Pearson correlations were computed to assess the correlation of the EEG divergence scores to BAC values.
