Measuring the accuracy of lifestyle-based judgments on political affiliation

Can we infer a person's voting preferences from their lifestyle choices? Are certain characteristics, such as the type of car you drive or leisure activities, really associated with different political parties?