It has often been suggested that reference group effects in education work through the self-concept of pupils.At the same time, it has been shown that teachers compare students with their peers when making evaluations.This observation gives rise to an alternative explanation of reference group effects on educational attainment: When deciding whether or not to recommend a student for grammar school, teachers consider not only the individual student's performance and abilities but use the wider group of motivated students as a frame of reference. Thus, given the limited number of places available in grammar school, a student's chances of being recommended also depend on his or her standing in the performance distribution as well as on the number of competitors. Using choice experiments presenting groups of three to five students to secondary school teachers in Switzerland, we show how the individual probability for a grammar school recommendation depends on the size of the reference group. Furthermore, individual chances are especially affected by the number of other students in the group that the teacher perceives as fit for grammar school. Our findings point towards a competition effect: The more and the better the competitors, the smaller the individual chances. In this respect, it seems better to be a big fish in a small pond.