This research proposal tackles the yet unanswered question as to how audiovisual translation deals with norm-related linguistic variation in the bi-centric Dutch language area (comprising both the Netherlands and Flanders). More particularly, it is investigated (i) whether subtitlers in Flanders prefer standard or colloquial Belgian Dutch lexemes and constructions rather than General Standard Dutch lexemes and constructions (accepted in both Flanders and the Netherlands), and (ii) to what extent their choices differ from those made by translators of other text types and by non-translators. To achieve these goals, we will compile several lexical and syntactic profiles, i.e. pairs of variants, with one being the General Dutch variant (e.g. ‘een beroep doen op’) and the other being the Belgian Dutch variant (e.g. ‘beroep doen op’). Those profiles will be extracted from different corpora, containing interlingual and intralingual subtitles on the one hand and translated and non-translated written text genres on the other hand. The obtained data will be manually validated and annotated for several contextual parameters (e.g., genre/register, target audience, type of editorial control, source language, status of the original speaker, interlingual vs. intralingual translation, professional subtitles vs. fansubs), and subjected to multivariate statistical techniques (a.o. correspondence analysis).