How intoxicated are you? Investigating self and observer intoxication ratings in relation to blood alcohol...

The ability to accurately detect alcohol intoxication is an important skill for people who use these substances and for a variety of professions (e.g. policing, responsible service of alcohol). Previous studies have found that intoxicated people are generally poor at estimating their own intoxication (particularly at high blood alcohol concentration; BAC) and the relationship between perceived intoxication and BAC appears to flatten at higher BAC levels. Studies of observer ratings of other's intoxication have yielded mixed results. The aim of this study was to investigate both self- and observer-ratings of intoxication against BAC levels to determine whether we observe a similar flattening in intoxication for both self- and observer-rated intoxication.

Participants were 388 students interviewed outside three university events. Participants provided demographics, a rating of how intoxicated they felt (0 = not at all, 10 = very) and provided an alcohol breath test. Interviewers recorded a rating of how intoxicated the participant appeared to be.

A significant correlation was observed between self- and observer-ratings of intoxication (r = 0.802). We fitted our data with both linear and polynomial regressions. Polynomial regression accounted for more variance when predicting both self-rated intoxication (R²  = 0.50 vs. R²  = 0.40) and observer-rated intoxication from BAC (R²  = 0.58 vs. R²  = 0.52), suggesting a flattening for both intoxication ratings.

Both self-rated and observer-rated intoxication appeared to flatten at higher levels of BAC. This may be due to either tolerance or a 'ceiling effect' for observable signs of intoxication.