The relationship between ethanol intake and DSM-III-R alcohol dependence.

The purpose of this investigation was to quantify the relationship between ethanol consumption and DSM-III-R alcohol dependence using mathematical modeling techniques that allowed for the control of confounding and assessment of interaction. Although sex, education, ethnicity and marital status were not identified as actual confounders in the logistic regression model, the ethanol intake-dependence association was stronger among younger as opposed to older respondents. For 20 year olds, the average log odds for dependence increased .62 for each additional ounce of ethanol consumed daily, while the corresponding increase in risk among 60 year olds was .26. Implications of these results are discussed in terms of age differences in drinking patterns and differential social control of drinking behavior. Separate analyses in which aggregates of the alcohol dependence criteria served as outcome measures helped qualify the interpretation of the overall ethanol intake-dependence relationship. The need to examine components of global classifications of alcohol dependence using better operationalizations is highlighted.