Search   |  Advanced

Research Summary

Accuracy of “Last Occasion” Self-Reported Drinking in Young Adults

Many studies rely on self-report of alcohol consumption, which may lead to a biased estimate of use. To assess the accuracy of "last occasion" self-reported alcohol consumption, researchers conducted a field study wherein individuals aged 18–25 were asked to report their alcohol consumption 1–2 days after being observed by peer-based researchers at various nightlife locations. The relationship between observed and reported consumption was assessed using multilevel models (129 observations for 48 individuals).

  • Overall, participants reported 9% fewer drinks than they actually drank.
  • There was a nonlinear relationship between reported and actual consumption:
    • Individuals drinking >8 drinks underestimated their consumption by increasing amounts (for example, those who consumed 12 drinks underestimated by 10.3% (1.3 drinks), while those who consumed 20 drinks underestimated by 17.6% (3.5 drinks).
    • Individuals engaging in less drinking (i.e., those who reported drinking ≤4 drinks or 5–8 drinks) accurately estimated their consumption.
  • Venue type did not impact the accuracy of self-report.


This study brings into question the accuracy of self-report measures used in population surveys, especially since reporting "last consumption" is easier than reporting "usual consumption," which requires individuals not only to recall past drinking accurately but also to estimate a "usual mean." It is, therefore, likely that the underestimation of self-reported drinking among individuals who drank heavily in this study will be also be found, and possibly to a greater extent, in general population surveys. Nicolas Bertholet, MD, MSc


Northcote J, Livingston M. Accuracy of self-reported drinking: observational verification of ‘last occasion’ drink estimates of young adults. Alcohol Alcohol. 2011;46(6):709–713.