Investigating Gender Differences in Alcohol Problems: A Latent Trait Modeling Approach

Authors


  • The research reported here was supported in part by USPHS Grant MH65137 and NIH Grants DA05147 and AA09367.

Reprint requests: Penny E. Nichol, BS, University of Minnesota, 75 East River Road, Minneapolis, MN 55455; Fax: 612-626-2079; E-mail: nich0185@umn.edu

Abstract

Background: Inconsistent results have been found in research investigating gender differences in alcohol problems. Previous studies of gender differences used a wide range of methodological techniques, as well as limited assortments of alcohol problems.

Methods: Parents (1,348 men and 1,402 women) of twins enrolled in the Minnesota Twin Family Study answered questions about a wide range of alcohol problems. A latent trait modeling technique was used to evaluate gender differences in the probability of endorsement at the problem level and for the overall 105-problem scale.

Results: Of the 34 problems that showed significant gender differences, 29 were more likely to be endorsed by men than women with equivalent overall alcohol problem levels. These male-oriented symptoms included measures of heavy drinking, duration of drinking, tolerance, and acting out behaviors. Nineteen symptoms were denoted for removal to create a scale that favored neither gender in assessment.

Conclusions: Significant gender differences were found in approximately one-third of the symptoms assessed and in the overall scale. Further examination of the nature of gender differences in alcohol problem symptoms should be undertaken to investigate whether a gender-neutral scale should be created or if men and women should be assessed with separate criteria for alcohol dependence and abuse.

Ancillary