This study compares the frequency and distribution of different types of corrective feedback (CF) (recasts, prompts and explicit correction) and learner uptake in 43 hours of classroom interaction at the 4th-5th grade level across three instructional settings: (1) two content and language integrated learning (CLIL) classrooms in Spain with English as the target language; (2) four French immersion (FI) classrooms in Quebec (using published data from Lyster and Ranta 1997); and (3) three Japanese immersion (JI) classrooms in the US (using published data from Lyster and Mori 2006). The findings revealed that teachers in all three settings used recasts, prompts and explicit correction in similar proportions, with recasts being the most frequent, followed by prompts then explicit correction. However, in the CLIL and JI classrooms, the majority of learner repair moves followed recasts, whereas the majority of repair moves in FI classrooms followed prompts. The similarities and differences across contexts are discussed in terms of the different types of recasts used by the teachers (i.e. didactic recasts in CLIL and JI; conversational recasts in FI), as well as in terms of context-specific influences and the teachers' professional trajectories.