Small-group interventions allow for tailored instruction for students with learning difficulties. A crucial first step is the accurate identification of students who need such an intervention. This study investigated how teachers decide whether their students need a remedial reading intervention. To this end, 64 teachers of 697 third-grade students from Germany were asked to rate whether a reading intervention for their students was "not necessary," "potentially necessary," or "definitely necessary." Independent experimenters tested the students' reading and spelling abilities with standardized tests, and a subsample of 370 children participated in standardized tests of phonological awareness and vocabulary. Findings show that teachers' decisions with regard to students' "needing" a reading intervention overlapped more with results from standardized spelling assessments than from reading assessments. Hierarchical linear models indicated that students' spelling abilities, along with phonological awareness and vocabulary, explained variance in teachers' ratings over and above students' reading skills. Teachers thus relied on proximal cues such as spelling skills to reach their decision. These findings are discussed in relation to clinical standards and educational contexts. Findings indicate that the teachers' assignment of children to interventions might be underspecified, and starting points for specific teacher training programs are outlined.