INTRODUCTION AND AIMS:Breathalyser estimate of blood alcohol concentration (BAC) is widely used as an objective intoxication measure, but is not always practical in nightlife contexts. This study uses in situ data collected in nightlife environments to explore how four measures of intoxication are related so as to inform the development of a more practical and reliable method of differentiating intoxication for people working in the night-time economy. DESIGN AND METHODS:Nightlife patron interviews were conducted in five Australian cities. Participants completed demographic questions and were asked about current session (past 12 h) alcohol use, and four different measures of intoxication were assessed: BAC, participant's self-reported intoxication (0-10), interviewer rating of the participant's intoxication (0-10) and interviewer-rated number of the participants' of physical signs of intoxication. RESULTS:A total of 7028 patrons were surveyed and n = 5273 included in analysis. Mean age was 23.9 years (SD = 6.36); 61.5% were male. There was a significant difference in occurrence of all observable intoxication symptoms across differing levels of BAC (P < 0.001). All visible symptoms became more common as intoxication increased, except for talking very quickly/talkative and giggly symptoms. As BAC levels increase, the extent of the disagreement between self-rated and interviewer-rated intoxication measures widens. Exhibiting four or more visible intoxication symptoms emerged as a reliable method for observers to identify intoxicated patrons. DISCUSSION AND CONCLUSIONS:As BAC increases, people become worse at estimating their own intoxication, but sober observers remain relatively accurate. Findings provide support for efforts to strengthen and enforce responsible service of alcohol.