This project focuses on the stylistic revision of Dutch–English translations. It investigates the effects of corpus-focused instruction (CFI) on stylistic translation revision competence. The use of corpora for translation revision has gained momentum in the last two decades. However, it remains unclear which corpora provide the best fit for the needs of translation trainees. Because error correction cannot take place without error detection, a crucial initial role in the process of stylistic translation revision is ascribed to receptive sensitivity to errors in translation revision tasks. However, detection of stylistic errors is difficult since it generally implies making choices between existing language variants with different degrees of formality. Language learners – including translation trainees – are often unaware that such choices must be made and if they are aware of such choices, they often find them extremely difficult. Consequently, an important question is how we can optimally develop receptive sensitivity to linguistic formality in translation settings to facilitate more accurate and informed style-related decisions during the translation revision process. This project addresses that question by investigating (1) how the form of explicit CFI affects translation trainees' ability to detect and correct formality-related problems in translation revision tasks and (2) which form of CFI triggers the highest success rate for detecting and correcting formality-related problems. The main variable in our experiment is corpus-focused instruction (prepared vs do-it-yourself corpora). Using these two types of corpora, we will investigate whether translation trainees' abundant contact with informal English affects their perception of linguistic formality and their ability to construct and/or use corpora. The results of this intervention will contribute to an informed best practice for translation revision pedagogy by defining a suggested degree of stylistic variation in corpora used for this purpose. Furthermore, the findings have theoretical significance for the fields of instructed second language acquisition, corpus-based translation studies and communication accommodation theory.