Abstrakt
In this paper, we are interested in whether the likelihood of an individual being an alcohol abuser can be predicted by some of descriptive socio-demographic factors such as age, marital status, smoking habits, education, employment status, income and municipality size. Unlike for other studies, which tend to assess the importance of an individual predictor disregarding possible influence of other individual characteristics, we let all descriptive variables enter a single logistic regression.
Only this way, we can properly assess their individual strength in defining propensity an individual's propensity to heavy binge drinking.