Figure 2

Results of ROC analysis (receiver operating characteristic). The ROC curve (black) describes the BIAT’s performance in discriminating dopers from non-dopers. The ROC graph depicts relative tradeoffs between true positives (test sensitivity) and false positives (test specificity) as a function of each obtained BIAT score. The farther the AUC (Area Under the Curve = 0.722 ± CI95%) deviates from the diagonal (grey) representing chance level (AUC = 0.50), the better the BIAT’s performance is in terms of maximizing true positive and true negative cases. At the proposed BIAT cut-off score (-0.57), specificity (54.3%) and sensitivity (84.6%) are maximized.