Measuring Teachers' Stereotypes in the NEPS

German data on the processes underlying discrimination is still sparse. Against this backdrop, this working paper discusses the measurement of a major source of discrimination: Stereotypes. Conceptualizing stereotypes as beliefs or sets of beliefs about the characteristics, attributes, or behaviors of a particular group of people, we introduce an explicit measure of teachers' stereotypes at the National Educational Panel Study (NEPS). Teachers are asked to estimate the average competencies of girls, boys, students of lower, middle, and upper class background, students of Turkish and Russian origin as well as immigrant students and ethnic majority students. Quantitative analyses based on a sample of 52 second-grade teachers show both a large variation in the expressed stereotypes between groups of students and between teachers. Furthermore, the analyses suggest that teachers' stereotypes are quite accurate overall in that they reflect group differences in achievement as reported in the recent literature. However, we also find biases to the disadvantage of boys, immigrants in general, as well as immigrants of Turkish and Russian origin in particular. We argue that these results and evidence from cognitive interviews speak for the validity of the instrument