The measurement of problem drinking in young adulthood.

Objective: Researchers and clinicians have been struggling for more than 20 years to define problem drinking. This study focuses on young adults and aims to examine the construct validity of three dimensions of problem drinking and to empirically determine appropriate cutpoints for problem drinking along these dimensions. Method: A stratified sample of 1,269 young adults who reported drinking alcohol in the year prior to 1995 was used in these analyses. Respondents were originally interviewed in 1985 in middle schools in a southeastern U.S. county. Results: Symptoms of dependency and drunkenness were relatively common in this sample, but adverse consequences were rare. Results of analyses using the Chi-squared Automatic Interaction Detector (CHAID) and the SUrvey DAta ANalysis (SUDAAN) software suggest that drunkenness and adverse consequences are the most significant predictors of problems in other areas of life. Symptoms of dependency were significant only in interaction with drunkenness, although the analyses were limited to only lifetime measures of dependency. Appropriate cutpoints, based on these analyses, appear to be (1) drunk at least six times in the past year, (2) four or more lifetime symptoms of dependency and (3) one or perhaps two or more adverse consequences in the past year. Conclusions: Results suggest that frequency of drunkenness may be the single best indicator of problem drinking among young adults and that adverse consequences may indicate a more serious form of problem drinking than do symptoms of dependency. In addition, appropriate cutpoints on these dimensions for young adults appear to be similar to those that have been used in studies of adolescents. Further study of both adolescents and young adults is suggested.