In two field studies, bar tenants ("N"s = 86 and 190, respectively) were successively approached by confederates C1 and C2 on a night out. Confederate C3 then presented participants with a six-person target-absent or target-present lineup concerning C1 or C2 (immediate test). Several days later, participants viewed a lineup regarding the confederate they had not attempted to identify earlier (C1/C2; delayed test). An immediate compared with a delayed, sober identification test did not increase the risk of a false identification decision. A blood alcohol concentration of 0.06-0.07% best discriminated accurate from inaccurate decisions. Choosers with a blood alcohol concentration [less than or equal to] 0.06% showed excellent calibration and little overconfidence, and their confidence was a strong indicator of accuracy (i.e., good resolution). Choosers with a higher intoxication level displayed poor calibration and strong overconfidence. Nonchoosers were generally poorly calibrated. Combined analyses showed a negative effect of intoxication on one's ability to discriminate the target from nontargets.