This study examines the impact of corpus-based feedback on elementary school students’ English writing performance and attitudes. Forty-two fifth-grade students were assigned to either a research group (n = 22), which received data-driven learning (DDL) feedback using teacher-selected concordance examples, or a comparison group (n = 20), which received direct teacher feedback and explicit instruction. Across four instructional lessons, both groups participated in pre- and post-writing assessments, and qualitative data were collected through surveys and interviews. The results indicate that the corpus-based group demonstrated greater improvement in writing performance compared to the comparison group, with significant gains observed among both higher- and lower-proficiency learners when appropriate scaffolding was provided. While students generally responded positively to corpus-based feedback, some reported difficulties during the writing process. These f indings suggest that corpus-based feedback can be an effective pedagogical approach in elementary writing instruction, offering a valuable alternative to conventional feedback methods.