Using Progress in International Reading Literacy Study 2001 and Programme for International Student Assessment 2003 data for Germany, this paper examines whether second-generation immigrants and girls are graded worse in math than comparable natives and boys, respectively. Once all grading-relevant characteristics, namely math skills and oral participation, are accounted for, pupils should obtain same school grades. Results of a matching approach and class fixed effects regressions suggest that second-generation immigrants have grade disadvantages in primary education which could bias their secondary school track choice. Regarding secondary school, most immigrants are not affected by grade discrimination and girls enro lled in upper-secondary school are systematically graded better.