This article examines whether teacher discrimination is an additional source of disadvantage for lower class and minority children, while also considering differential treatment by gender. Primary teachers in Bern, Switzerland, were asked to state the probabilities of recommending fictitious sixth graders – described in vignettes – to either the lower or higher track in secondary school. The factorial survey experiment, especially suited for addressing problems of social desirability bias, while enabling a direct measure of attitudes, included the dimensions: social class, minority status, gender, ability, motivation, classroom behaviour and parental educational aspirations. The results from the ordinary least squares regression models, accounting for the hierarchical structure (each respondent evaluated four vignettes so that N=216) with clustered standard errors, were mixed. While teachers lean towards treating ethnic minority girls favourably, they are significantly harsher on minority boys, compared to their Swiss counterparts. Throughout, teachers respond somewhat differently to girls and boys; sex-specific effects emerge for motivation, classroom behaviour and parental educational aspirations. In regard to social class, teachers tend towards a downward bias for lower class children in general, but it is yet again the boys that have a substantial disadvantage if they come from a lower class background. On the whole, the results indicate that discriminatory behaviour by teachers may indeed be one of the reasons for the lower school achievement of lower class and minority boys. While the theories of preference-based and statistical discrimination fail to account for the findings, the theory of statistical discrimination proves valuable in explaining the results.