Self-concept in mathematics (MSC) and interest in mathematics are important predictors of whether a student will choose to major in a science, technology, engineering, or mathematics (STEM) field at university. Research on the big-fish-little-pond effect (BFLPE) has shown that both predictors are affected by the achievement composition of students in schools. That is, given the same individual ability, students in higher achieving schools exhibit lower MSC and interest in mathematics than students in lower achieving schools due to social comparison processes. The BFLPE has been replicated in high school settings many times, but less is known about the long-term effects of this context effect. In the present study, we investigated the long-term effects of the BFLPE on the aspiration to major in and the actual decision to major in a STEM field at university. We used data from a German longitudinal study. The results showed no direct BFLPE on the aspiration to enroll in a STEM major at the end of high school or on actual enrollment in a STEM major 2 years after graduating from high school. However, small negative indirect effects of the BFLPE via MSC and interest in mathematics on the aspiration to and enrollment in a STEM major occurred. In sum, the longitudinal BFLPE on STEM major choice was small.