In a reference grammar of English for Danish students, Hjulmand & Schwarz (2015: 137) state that, when translating from Danish, “'en hel del' is a good/great deal of in front of uncountable nouns, but a good/great many in front of countable nouns in the plural”.This claim calls for empirical support. With significant distributions of countnouns vs. non-count nouns, a study of COCA suggests that the claim holds up atleast for American English. However, the claim ultimately belongs to what Harder (2015; see also Gregory 1967) calls incomplete accounts. In the perspectiveof usage-based linguistics, such a claim would leave out informationpotentially useful to Danish learners of English. Drawing on principles fromconstruction grammar (e.g. Goldberg 1995; Croft 2001) and variationist cognitive sociolinguistics (Pütz et al. 2014), this paper presents a usage-based comparativecorpus study of the two constructions. Drawing on data from COCA, a distinctivecollexeme analysis (Gries & Stefanowitsch 2004) shows that, not only do the constructionsdiffer in terms of preference for count vs. non-count nouns, they also havedifferent preferences for specific individual nouns and semantic classes ofnouns. Moreover, variety-centered multidimensional scaling analyses andheatmaps indicate that the patterns of use of the constructions displayregister variation. In addition, a lexical richness analysis revealsdifferences in constructional productivity.