Send to

Choose Destination
J Stud Alcohol. 1997 Mar;58(2):167-81.

Predictors of problem drinking and alcohol dependence in a population-based sample of female twins.

Author information

Department of Psychiatry, Virginia Institute for Psychiatric and Behavioral Genetics, Medical College of Virginia, Virginia Commonwealth University, Richmond 23298-0124, USA.



To identify characteristics associated with problem drinking (PD) and alcohol dependence (AD) in women.


Subjects were 2,163 white women aged 17-55 from the population-based Virginia Twin Registry. Measures were selected from a clinical interview and questionnaires to reflect five domains associated with alcoholism in prior studies; demographic characteristics, personality, health, and personal and family history of psychopathology. Logistic and linear regression analyses were used to predict PD and DSM-III-R defined AD.


Multiple regression models accounted for 19% of the variance in PD (significant predictors included: higher parental education-particularly among younger women, being the primary breadwinner, less frequent church attendance, higher scores on measures of neuroticism, extraversion and interpersonal dependency, history of major depression and social phobia, paternal PD and maternal treatment for emotional problems); 9% of the variance in diagnosis of AD (predicted by generalized anxiety, paternal depression and maternal PD); and 20% of the variance in number of symptoms of AD (predicted by the interaction of younger age and less-educated parents, higher neuroticism and mastery, lower optimism, generalized anxiety and agoraphobia, and maternal PD).


Personality characteristics and parental psychopathology are important predictors of PD and AD independent of their effect on risk for affective and anxiety disorders. Many characteristics found to be associated with PD and AD in bivariate analyses were not significant when considered in the context of other predictors. Future studies of the etiology of alcoholism among women should simultaneously study measures from a variety of domains.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Dartmouth Journal Services
Loading ...
Support Center