This paper presents the results of a conjoint survey experiment in which Swiss citizens were asked to choose among parliamentary candidates with different class profiles determined by occupation, education and income. Existing survey-experimental literature on this topic suggests that respondents are indifferent to the class profiles of candidates or biased against candidates with high-status occupations and high incomes. We find that respondents are biased against upper middle-class candidates as well as routine working-class candidates. While the bias against upper middle-class candidates is primarily a bias among working-class individuals, the bias against routine working-class candidates is most pronounced among middle-class individuals. Our supplementary analysis of observational data confirms the bias against routine working-class candidates, but not the bias against upper middle-class candidates.