Teachers are often asked to nominate students for enrichment programs for gifted children, and studies have repeatedly indicated that students' intelligence is related to their likelihood of being nominated as gifted. However, it is unknown whether class-average levels of intelligence influence teachers' nominations as suggested by theory--and corresponding empirical results--concerning reference group effects. Herein, it was hypothesized that when students' individual fluid and crystallized intelligence scores were similar, students from classes with higher average levels of intelligence would have a lower probability of being nominated for an enrichment program for gifted children than students from classes with lower average levels of intelligence. Furthermore, we investigated whether 3 teacher variables--experience with giftedness, beliefs about the changeability of intelligence, and the belief that giftedness is holistic or domain specific--would influence the expected reference group effect. In a study comprising data from 105 teachers and 1,468 of their (German) third-grade students, we found support not only for a positive association between students' individual intelligence scores and the probability that students would be nominated as gifted but also, more importantly, for the proposed reference group effect: When controlling for individual levels of intelligence, students' probability of being nominated was higher in classes with lower average levels of intelligence. In addition, the results showed that this reference group effect was stronger when teachers saw giftedness as holistic rather than domain specific. Also, depending on teachers' kinds of experience with giftedness, the reference group effect varied in size.