Environmental and individual predictors of error in field estimates of blood alcohol concentration: A multilevel analysis

Objective: Using self-report survey data and measures of breath alcohol concentration (BrAC), this study examined the validity of an estimate of blood alcohol concentration (eBAC). Differences between eBAC and BrAC were modeled to identify sources of error related to survey-derived eBAC. Further, using hierarchical multilevel analyses, environmental event characteristics were examined as sources of eBAC error. Method: College students were interviewed and provided breath samples at randomly selected parties on Friday and Saturday nights. Interviews included items assessing the total number of drinks consumed, duration of drinking event, gender, and weight, which allowed for the calculation of eBAC. Results: Overall, eBACs were inaccurate. Total number of drinks consumed was associated with underestimates of eBAC, whereas time drinking was associated with overestimates of eBAC. Environmental variables, including party size, rowdy behavior, having food present, and observing many intoxicated partygoers, were also associated with eBAC errors. Conclusions: Current self-report survey methodology to calculate eBAC may be insufficient to estimate BAC with any accuracy. Environmental factors associated with the last drinking event for which BAC is being estimated should be considered when calculating eBAC.