This article seems to be claiming that people with a larger vocabulary (as tested by a vocab test) are more likely to drink, not 'people of class'.
Summary of that comment: Attitudes towards scripture is the primary predictor, though WORDSUM and education still have significant effects.