Objective: This field study of late-night college drinking sought to (1) test the ability of the 5+/4+ measure to screen for higher levels of intoxication and (2) examine the relation between estimated and actual blood alcohol concentration (BAC). Method: During a 15-week spring semester, college students returning to their residence halls between 10:00 PM and 3:00 AM on Wednesday through Saturday nights were anonymously interviewed to collect BAC and self-report data (n = 1,020). Results: Although 70.9% had not been drinking on Wednesday nights, a majority of the intercepted students had been drinking on the other three nights. Mean BACs on these three nights were in a moderate range (48 to 51 mg/dl), but the 5+/4+ measure classified many students as heavy episodic drinkers at relatively low BACs. For example, 66.3% of those meeting the 5+/4+ criterion for the night had BACs < 100 mg/dl. Students with BACs ranging from 70 to 90 mg/dl exhibited the greatest accuracy in estimating their BAC; those with lower BACs tended to overestimate their level of intoxication; whereas those with higher BACs tended to underestimate it. Conclusions: Field assessment of student intoxication is an important tool for examining research questions in college drinking. The 5+/4+ measure classifies many college students as heavy episodic drinkers, even though their intoxication level is below conventional thresholds used to define drunkenness. In addition, there is a discernible pattern of BAC estimation in the field that corresponds to intoxication level.