The current study uses a large, nationally representative data set and a new method for computing teacher expectations to better understand the developmental effect of mathematics teacher expectations on future student achievement. The study utilizes autoregressive cross-lagged models with 5 time points between kindergarten and 8th grade as well as multigroup modeling to examine group differences in teacher expectancy effects on achievement for girls and minority students. Results indicate that students' experiences with teacher expectations from 1 time point to the next are not significantly associated with one another, but their association with future student achievement grows over time. Teacher expectancy effects in mathematics are stronger for White girls, minority girls, and minority boys than they are for White boys. Implications for teaching are discussed.