This paper looks at the alternation of two complex English noun phrases in scientific English, which poses a challenge to French students in the specialised translation classroom. Indeed, no such alternation is observed in French. Starting from a preliminary study of a first series of constructions, we seek confirmation for generalisations about the constructions' preferred context of occurrence in a new sample of highly frequent constructions. We then discuss how the results of those analyses can be integrated in the translation classroom, through a new online tool aimed at raising students' awareness of this contrastive problem and helping them choose one or the other construction according to a set of corpus-based clues.