The aim of this paper is to account for a case study based on the Bergen Translation Corpus (BTC), a project launched to identify specific translation challenges in connection with the National Translator Accreditation Exam (NTAE) in Norway. The corpus is made up of translations by candidates entering the NTAE. The exam is held with no requirement to have previously attended a study programme in translation studies. The BTC is a parallel multilingual corpus, primarily compiled for research purposes, but subsequently used for training purposes of candidates in the context of the NTAE. This paper introduces a research-based and training-oriented application, with the aim to identify candidates' handling of two specific translation challenges in legal texts, viz. culturally bound concepts (cultural references) and modality. The small number of translations included in the BTC allows only for a qualitative analysis. Results of the case study support findings from previous studies on those issues.